- Pre-registration nursing students
- No definition of master’s degree in nursing described in the publication
After the search, we collated and uploaded all the identified records into EndNote v.X8 (Clarivate Analytics, Philadelphia, Pennsylvania) and removed any duplicates. Two independent reviewers (MCS and SA) screened the titles and abstracts for assessment in line with the inclusion criteria. They retrieved and assessed the full texts of the selected studies while applying the inclusion criteria. Any disagreements about the eligibility of studies were resolved by discussion or, if no consensus could be reached, by involving experienced researchers (MZ-S and RP).
The first reviewer (MCS) extracted data from the selected publications. For this purpose, an extraction tool developed by the authors was used. This tool comprised the following criteria: author(s), year of publication, country, research question, design, case definition, data sources, and methodologic and data-analysis triangulation. First, we extracted and summarized information about the case study design. Second, we narratively summarized the way in which the data and methodological triangulation were described. Finally, we summarized the information on within-case or cross-case analysis. This process was performed using Microsoft Excel. One reviewer (MCS) extracted data, whereas another reviewer (SA) cross-checked the data extraction, making suggestions for additions or edits. Any disagreements between the reviewers were resolved through discussion.
A total of 149 records were identified in 2 databases. We removed 20 duplicates and screened 129 reports by title and abstract. A total of 46 reports were assessed for eligibility. Through hand searches, we identified 117 additional records. Of these, we excluded 98 reports after title and abstract screening. A total of 17 reports were assessed for eligibility. From the 2 databases and the hand search, 63 reports were assessed for eligibility. Ultimately, we included 8 articles for data extraction. No further articles were included after the reference list screening of the included studies. A PRISMA flow diagram of the study selection and inclusion process is presented in Figure 1 . As shown in Tables 2 and and3, 3 , the articles included in this scoping review were published between 2010 and 2022 in Canada (n = 3), the United States (n = 2), Australia (n = 2), and Scotland (n = 1).
PRISMA flow diagram.
Characteristics of Articles Included.
Author | Contandriopoulos et al | Flinter | Hogan et al | Hungerford et al | O’Rourke | Roots and MacDonald | Schadewaldt et al | Strachan et al |
---|---|---|---|---|---|---|---|---|
Country | Canada | The United States | The United States | Australia | Canada | Canada | Australia | Scotland |
How or why research question | No information on the research question | Several how or why research questions | What and how research question | No information on the research question | Several how or why research questions | No information on the research question | What research question | What and why research questions |
Design and referenced author of methodological guidance | Six qualitative case studies Robert K. Yin | Multiple-case studies design Robert K. Yin | Multiple-case studies design Robert E. Stake | Case study design Robert K. Yin | Qualitative single-case study Robert K. Yin Robert E. Stake Sharan Merriam | Single-case study design Robert K. Yin Sharan Merriam | Multiple-case studies design Robert K. Yin Robert E. Stake | Multiple-case studies design |
Case definition | Team of health professionals (Small group) | Nurse practitioners (Individuals) | Primary care practices (Organization) | Community-based NP model of practice (Organization) | NP-led practice (Organization) | Primary care practices (Organization) | No information on case definition | Health board (Organization) |
Overview of Within-Method, Between/Across-Method, and Data-Analysis Triangulation.
Author | Contandriopoulos et al | Flinter | Hogan et al | Hungerford et al | O’Rourke | Roots and MacDonald | Schadewaldt et al | Strachan et al |
---|---|---|---|---|---|---|---|---|
Within-method triangulation (using within-method triangulation use at least 2 data-collection procedures from the same design approach) | ||||||||
: | ||||||||
Interviews | X | x | x | x | x | |||
Observations | x | x | ||||||
Public documents | x | x | x | |||||
Electronic health records | x | |||||||
Between/across-method (using both qualitative and quantitative data-collection procedures in the same study) | ||||||||
: | ||||||||
: | ||||||||
Interviews | x | x | x | |||||
Observations | x | x | ||||||
Public documents | x | x | ||||||
Electronic health records | x | |||||||
: | ||||||||
Self-assessment | x | |||||||
Service records | x | |||||||
Questionnaires | x | |||||||
Data-analysis triangulation (combination of 2 or more methods of analyzing data) | ||||||||
: | ||||||||
: | ||||||||
Deductive | x | x | x | |||||
Inductive | x | x | ||||||
Thematic | x | x | ||||||
Content | ||||||||
: | ||||||||
Descriptive analysis | x | x | x | |||||
: | ||||||||
: | ||||||||
Deductive | x | x | x | x | ||||
Inductive | x | x | ||||||
Thematic | x | |||||||
Content | x |
The following sections describe the research question, case definition, and case study design. Case studies are most appropriate when asking “how” or “why” questions. 1 According to Yin, 1 how and why questions are explanatory and lead to the use of case studies, histories, and experiments as the preferred research methods. In 1 study from Canada, eg, the following research question was presented: “How and why did stakeholders participate in the system change process that led to the introduction of the first nurse practitioner-led Clinic in Ontario?” (p7) 19 Once the research question has been formulated, the case should be defined and, subsequently, the case study design chosen. 1 In typical case studies with mixed methods, the 2 types of data are gathered concurrently in a convergent design and the results merged to examine a case and/or compare multiple cases. 10
“How” or “why” questions were found in 4 studies. 16 , 17 , 19 , 22 Two studies additionally asked “what” questions. Three studies described an exploratory approach, and 1 study presented an explanatory approach. Of these 4 studies, 3 studies chose a qualitative approach 17 , 19 , 22 and 1 opted for mixed methods with a convergent design. 16
In the remaining studies, either the research questions were not clearly stated or no “how” or “why” questions were formulated. For example, “what” questions were found in 1 study. 21 No information was provided on exploratory, descriptive, and explanatory approaches. Schadewaldt et al 21 chose mixed methods with a convergent design.
A total of 5 studies defined the case as an organizational unit. 17 , 18 - 20 , 22 Of the 8 articles, 4 reported multiple-case studies. 16 , 17 , 22 , 23 Another 2 publications involved single-case studies. 19 , 20 Moreover, 2 publications did not state the case study design explicitly.
This section describes within-method triangulation, which involves employing at least 2 data-collection procedures within the same design approach. 6 , 7 This can also be called data source triangulation. 8 Next, we present the single data-collection procedures in detail. In 5 studies, information on within-method triangulation was found. 15 , 17 - 19 , 22 Studies describing a quantitative approach and the triangulation of 2 or more quantitative data-collection procedures could not be included in this scoping review.
Five studies used qualitative data-collection procedures. Two studies combined face-to-face interviews and documents. 15 , 19 One study mixed in-depth interviews with observations, 18 and 1 study combined face-to-face interviews and documentation. 22 One study contained face-to-face interviews, observations, and documentation. 17 The combination of different qualitative data-collection procedures was used to present the case context in an authentic and complex way, to elicit the perspectives of the participants, and to obtain a holistic description and explanation of the cases under study.
All 5 studies used qualitative interviews as the primary data-collection procedure. 15 , 17 - 19 , 22 Face-to-face, in-depth, and semi-structured interviews were conducted. The topics covered in the interviews included processes in the introduction of new care services and experiences of barriers and facilitators to collaborative work in general practices. Two studies did not specify the type of interviews conducted and did not report sample questions. 15 , 18
In 2 studies, qualitative observations were carried out. 17 , 18 During the observations, the physical design of the clinical patients’ rooms and office spaces was examined. 17 Hungerford et al 18 did not explain what information was collected during the observations. In both studies, the type of observation was not specified. Observations were generally recorded as field notes.
In 3 studies, various qualitative public documents were studied. 15 , 19 , 22 These documents included role description, education curriculum, governance frameworks, websites, and newspapers with information about the implementation of the role and general practice. Only 1 study failed to specify the type of document and the collected data. 15
In 1 study, qualitative documentation was investigated. 17 This included a review of dashboards (eg, provider productivity reports or provider quality dashboards in the electronic health record) and quality performance reports (eg, practice-wide or co-management team-wide performance reports).
This section describes the between/across methods, which involve employing both qualitative and quantitative data-collection procedures in the same study. 6 , 7 This procedure can also be denoted “methodologic triangulation.” 8 Subsequently, we present the individual data-collection procedures. In 3 studies, information on between/across triangulation was found. 16 , 20 , 21
Three studies used qualitative and quantitative data-collection procedures. One study combined face-to-face interviews, documentation, and self-assessments. 16 One study employed semi-structured interviews, direct observation, documents, and service records, 20 and another study combined face-to-face interviews, non-participant observation, documents, and questionnaires. 23
All 3 studies used qualitative interviews as the primary data-collection procedure. 16 , 20 , 23 Face-to-face and semi-structured interviews were conducted. In the interviews, data were collected on the introduction of new care services and experiences of barriers to and facilitators of collaborative work in general practices.
In 2 studies, direct and non-participant qualitative observations were conducted. 20 , 23 During the observations, the interaction between health professionals or the organization and the clinical context was observed. Observations were generally recorded as field notes.
In 2 studies, various qualitative public documents were examined. 20 , 23 These documents included role description, newspapers, websites, and practice documents (eg, flyers). In the documents, information on the role implementation and role description of NPs was collected.
In 1 study, qualitative individual journals were studied. 16 These included reflective journals from NPs, who performed the role in primary health care.
Only 1 study involved quantitative service records. 20 These service records were obtained from the primary care practices and the respective health authorities. They were collected before and after the implementation of an NP role to identify changes in patients’ access to health care, the volume of patients served, and patients’ use of acute care services.
In 2 studies, quantitative questionnaires were used to gather information about the teams’ satisfaction with collaboration. 16 , 21 In 1 study, 3 validated scales were used. The scales measured experience, satisfaction, and belief in the benefits of collaboration. 21 Psychometric performance indicators of these scales were provided. However, the time points of data collection were not specified; similarly, whether the questionnaires were completed online or by hand was not mentioned. A competency self-assessment tool was used in another study. 16 The assessment comprised 70 items and included topics such as health promotion, protection, disease prevention and treatment, the NP-patient relationship, the teaching-coaching function, the professional role, managing and negotiating health care delivery systems, monitoring and ensuring the quality of health care practice, and cultural competence. Psychometric performance indicators were provided. The assessment was completed online with 2 measurement time points (pre self-assessment and post self-assessment).
This section describes data-analysis triangulation, which involves the combination of 2 or more methods of analyzing data. 6 Subsequently, we present within-case analysis and cross-case analysis.
Three studies combined qualitative and quantitative methods of analysis. 16 , 20 , 21 Two studies involved deductive and inductive qualitative analysis, and qualitative data were analyzed thematically. 20 , 21 One used deductive qualitative analysis. 16 The method of analysis was not specified in the studies. Quantitative data were analyzed using descriptive statistics in 3 studies. 16 , 20 , 23 The descriptive statistics comprised the calculation of the mean, median, and frequencies.
Two studies combined deductive and inductive qualitative analysis, 19 , 22 and 2 studies only used deductive qualitative analysis. 15 , 18 Qualitative data were analyzed thematically in 1 study, 22 and data were treated with content analysis in the other. 19 The method of analysis was not specified in the 2 studies.
In 7 studies, a within-case analysis was performed. 15 - 20 , 22 Six studies used qualitative data for the within-case analysis, and 1 study employed qualitative and quantitative data. Data were analyzed separately, consecutively, or in parallel. The themes generated from qualitative data were compared and then summarized. The individual cases were presented mostly as a narrative description. Quantitative data were integrated into the qualitative description with tables and graphs. Qualitative and quantitative data were also presented as a narrative description.
Of the multiple-case studies, 5 carried out cross-case analyses. 15 - 17 , 20 , 22 Three studies described the cross-case analysis using qualitative data. Two studies reported a combination of qualitative and quantitative data for the cross-case analysis. In each multiple-case study, the individual cases were contrasted to identify the differences and similarities between the cases. One study did not specify whether a within-case or a cross-case analysis was conducted. 23
This section describes confirmation or contradiction through qualitative and quantitative data. 1 , 4 Qualitative and quantitative data were reported separately, with little connection between them. As a result, the conclusions on neither the comparisons nor the contradictions could be clearly determined.
In 3 studies, the consistency of the results of different types of qualitative data was highlighted. 16 , 19 , 21 In particular, documentation and interviews or interviews and observations were contrasted:
Both types of data showed that NPs and general practitioners wanted to have more time in common to discuss patient cases and engage in personal exchanges. 21 In addition, the qualitative and quantitative data confirmed the individual progression of NPs from less competent to more competent. 16 One study pointed out that qualitative and quantitative data obtained similar results for the cases. 20 For example, integrating NPs improved patient access by increasing appointment availability.
Although questionnaire results indicated that NPs and general practitioners experienced high levels of collaboration and satisfaction with the collaborative relationship, the qualitative results drew a more ambivalent picture of NPs’ and general practitioners’ experiences with collaboration. 21
The studies included in this scoping review evidenced various research questions. The recommended formats (ie, how or why questions) were not applied consistently. Therefore, no case study design should be applied because the research question is the major guide for determining the research design. 2 Furthermore, case definitions and designs were applied variably. The lack of standardization is reflected in differences in the reporting of these case studies. Generally, case study research is viewed as allowing much more freedom and flexibility. 5 , 24 However, this flexibility and the lack of uniform specifications lead to confusion.
Methodologic triangulation, as described in the literature, can be somewhat confusing as it can refer to either data-collection methods or research designs. 6 , 8 For example, methodologic triangulation can allude to qualitative and quantitative methods, indicating a paradigmatic connection. Methodologic triangulation can also point to qualitative and quantitative data-collection methods, analysis, and interpretation without specific philosophical stances. 6 , 8 Regarding “data-collection methods with no philosophical stances,” we would recommend using the wording “data source triangulation” instead. Thus, the demarcation between the method and the data-collection procedures will be clearer.
Yin 1 advocated the use of multiple sources of evidence so that a case or cases can be investigated more comprehensively and accurately. Most studies included multiple data-collection procedures. Five studies employed a variety of qualitative data-collection procedures, and 3 studies used qualitative and quantitative data-collection procedures (mixed methods). In contrast, no study contained 2 or more quantitative data-collection procedures. In particular, quantitative data-collection procedures—such as validated, reliable questionnaires, scales, or assessments—were not used exhaustively. The prerequisites for using multiple data-collection procedures are availability, the knowledge and skill of the researcher, and sufficient financial funds. 1 To meet these prerequisites, research teams consisting of members with different levels of training and experience are necessary. Multidisciplinary research teams need to be aware of the strengths and weaknesses of different data sources and collection procedures. 1
When using multiple data sources and analysis methods, it is necessary to present the results in a coherent manner. Although the importance of multiple data sources and analysis has been emphasized, 1 , 5 the description of triangulation has tended to be brief. Thus, traceability of the research process is not always ensured. The sparse description of the data-analysis triangulation procedure may be due to the limited number of words in publications or the complexity involved in merging the different data sources.
Only a few concrete recommendations regarding the operationalization of the data-analysis triangulation with the qualitative data process were found. 25 A total of 3 approaches have been proposed 25 : (1) the intuitive approach, in which researchers intuitively connect information from different data sources; (2) the procedural approach, in which each comparative or contrasting step in triangulation is documented to ensure transparency and replicability; and (3) the intersubjective approach, which necessitates a group of researchers agreeing on the steps in the triangulation process. For each case study, one of these 3 approaches needs to be selected, carefully carried out, and documented. Thus, in-depth examination of the data can take place. Farmer et al 25 concluded that most researchers take the intuitive approach; therefore, triangulation is not clearly articulated. This trend is also evident in our scoping review.
Few studies in this scoping review used a combination of qualitative and quantitative analysis. However, creating a comprehensive stand-alone picture of a case from both qualitative and quantitative methods is challenging. Findings derived from different data types may not automatically coalesce into a coherent whole. 4 O’Cathain et al 26 described 3 techniques for combining the results of qualitative and quantitative methods: (1) developing a triangulation protocol; (2) following a thread by selecting a theme from 1 component and following it across the other components; and (3) developing a mixed-methods matrix.
The most detailed description of the conducting of triangulation is the triangulation protocol. The triangulation protocol takes place at the interpretation stage of the research process. 26 This protocol was developed for multiple qualitative data but can also be applied to a combination of qualitative and quantitative data. 25 , 26 It is possible to determine agreement, partial agreement, “silence,” or dissonance between the results of qualitative and quantitative data. The protocol is intended to bring together the various themes from the qualitative and quantitative results and identify overarching meta-themes. 25 , 26
The “following a thread” technique is used in the analysis stage of the research process. To begin, each data source is analyzed to identify the most important themes that need further investigation. Subsequently, the research team selects 1 theme from 1 data source and follows it up in the other data source, thereby creating a thread. The individual steps of this technique are not specified. 26 , 27
A mixed-methods matrix is used at the end of the analysis. 26 All the data collected on a defined case are examined together in 1 large matrix, paying attention to cases rather than variables or themes. In a mixed-methods matrix (eg, a table), the rows represent the cases for which both qualitative and quantitative data exist. The columns show the findings for each case. This technique allows the research team to look for congruency, surprises, and paradoxes among the findings as well as patterns across multiple cases. In our review, we identified only one of these 3 approaches in the study by Roots and MacDonald. 20 These authors mentioned that a causal network analysis was performed using a matrix. However, no further details were given, and reference was made to a later publication. We could not find this publication.
Because it focused on the implementation of NPs in primary health care, the setting of this scoping review was narrow. However, triangulation is essential for research in this area. This type of research was found to provide a good basis for understanding methodologic and data-analysis triangulation. Despite the lack of traceability in the description of the data and methodological triangulation, we believe that case studies are an appropriate design for exploring new nursing roles in existing health care systems. This is evidenced by the fact that case study research is widely used in many social science disciplines as well as in professional practice. 1 To strengthen this research method and increase the traceability in the research process, we recommend using the reporting guideline and reporting checklist by Rodgers et al. 9 This reporting checklist needs to be complemented with methodologic and data-analysis triangulation. A procedural approach needs to be followed in which each comparative step of the triangulation is documented. 25 A triangulation protocol or a mixed-methods matrix can be used for this purpose. 26 If there is a word limit in a publication, the triangulation protocol or mixed-methods matrix needs to be identified. A schematic representation of methodologic and data-analysis triangulation in case studies can be found in Figure 2 .
Schematic representation of methodologic and data-analysis triangulation in case studies (own work).
This study suffered from several limitations that must be acknowledged. Given the nature of scoping reviews, we did not analyze the evidence reported in the studies. However, 2 reviewers independently reviewed all the full-text reports with respect to the inclusion criteria. The focus on the primary care setting with NPs (master’s degree) was very narrow, and only a few studies qualified. Thus, possible important methodological aspects that would have contributed to answering the questions were omitted. Studies describing the triangulation of 2 or more quantitative data-collection procedures could not be included in this scoping review due to the inclusion and exclusion criteria.
Given the various processes described for methodologic and data-analysis triangulation, we can conclude that triangulation in case studies is poorly standardized. Consequently, the traceability of the research process is not always given. Triangulation is complicated by the confusion of terminology. To advance case study research in nursing, we encourage authors to reflect critically on methodologic and data-analysis triangulation and use existing tools, such as the triangulation protocol or mixed-methods matrix and the reporting guideline checklist by Rodgers et al, 9 to ensure more transparent reporting.
Acknowledgments.
The authors thank Simona Aeschlimann for her support during the screening process.
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding: The author(s) received no financial support for the research, authorship, and/or publication of this article.
Supplemental Material: Supplemental material for this article is available online.
You have full access to this open access article
Cultural data analytics aims to use analytic methods to explore cultural expressions—for instance art, literature, dance, music. The common thing between cultural expressions is that they have multiple qualitatively different facets that interact with each other in non trivial and non learnable ways. To support this observation, we use the Italian music record industry from 1902 to 2024 as a case study. In this scenario, a possible research objective could be to discuss the relationships between different music genres as they are performed by different bands. Estimating genre similarity by counting the number of records each band published performing a given genre is not enough, because it assumes bands operate independently from each other. In reality, bands share members and have complex relationships. These relationships cannot be automatically learned, both because we miss the data behind their creation, but also because they are established in a serendipitous way between artists, without following consistent patterns. However, we can be map them in a complex network. We can then use the counts of band records with a given genre as a node attribute in a band network. In this paper we show how recently developed techniques for node attribute analysis are a natural choice to analyze such attributes. Alternative network analysis techniques focus on analyzing nodes, rather than node attributes, ending up either being inapplicable in this scenario, or requiring the creation of more complex n-partite high order structures that can result less intuitive. By using node attribute analysis techniques, we show that we are able to describe which music genres concentrate or spread out in this network, which time periods show a balance of exploration-versus-exploitation, which Italian regions correlate more with which music genres, and a new approach to classify clusters of coherent music genres or eras of activity by the distance on this network between genres or years.
Node attribute analysis has recently been enlarged by the introduction of techniques to calculate the variance of a node attribute (Devriendt et al. 2022 ), estimate distances between two node attributes (Coscia 2020 ), calculating their Pearson correlations (Coscia 2021 ), and cluster them (Damstrup et al. 2023 ) without assuming they live in a simple Euclidean space—or learnable deformation thereof.
These techniques are useful only insofar the network being analyzed has rich node attribute data, and that analyzing their relationships is interesting. This is normally the case in cultural analytics, the use of analytic methods for the exploration of contemporary and historical cultures (Manovich 2020 ; Candia et al. 2019 ). Example range from archaeology—where related artifacts have a number of physical characteristics and can be from different places/ages (Schich et al. 2008 ; Brughmans 2013 ; Mills et al. 2013 ); to art history—where related visual artifacts can be described by a number of meaningful visual characteristics (Salah et al. 2013 ; Hristova 2016 ; Karjus et al. 2023 ); to sociology—where different ideas and opinions distribute over a social network as node attributes (Bail 2014 ; Hohmann et al. 2023 ); to linguistics—with different people in a social network producing content in different languages (Ronen et al. 2014 ); to music—with complex relations between players and informing meta-relationships between the genres they play (McAndrew and Everett 2015 ; Vlegels and Lievens 2017 ).
In this paper we aim at showing the usefulness of node attribute analysis in cultural analytics. We focus on the Italian record music industry since its beginnings in the early XX century until the present time. We build a temporally-evolving bipartite network connecting players with the bands they play in. For each band we know how many records of a given genre they publish, whether they published a record in a given year, and from which Italian region they originate—all node attributes of the band. By applying node attribute analysis, we can address a number of interesting questions. For instance:
How related is a particular music genre to a period? Or to a specific Italian region?
Is the production of a specific genre concentrated in a restricted group of bands or generally spread through the network?
Does clustering genres according to their distribution on the collaboration network conform to our expectation of meta-genres or can we discover a new network-based classification?
Can we use the productivity of related bands across the years as the basis to find eras in music production?
The music scene has been the subject of extensive analysis using networks. Some works focus on music production as an import–export network between countries (Moon et al. 2010 ). Other model composers and performers as nodes connected by collaboration or friendship links (Stebbins 2004 ; Park et al. 2007 ; Gleiser and Danon 2003 ; Teitelbaum et al. 2008 ; McAndrew and Everett 2015 ). Studies investigate how music consumption can inform us about genres (Vlegels and Lievens 2017 ) and listeners influencing each other (Baym and Ledbetter 2009 ; Pennacchioli et al. 2013 ; Pálovics and Benczúr 2013 ). Differently from these studies, we do not focus on asking questions about the network structure itself. For our work, the network structure is interesting only insofar it is the mediator of the relationships between node attributes—the genres, years, and regions the bands are active on –, rather than being the focus of the analysis.
This is an important qualitative distinction, because if one wanted to perform our genre-regional analysis on the music collaboration network without our node attribute analysis, they would have to deal with complex n-partite objects—a player-band-year-genre-region network—which can become unwieldy and unintuitive. On the other hand, with our approach one can work with a unipartite projection of the player-band relationships, and use years, genres, and regions as node attributes, maintaining a highly intuitive representation.
Deep learning techniques and specifically deep neural networks can handle the richness of our data (Aljalbout et al. 2018 ; Aggarwal et al. 2018 ; Pang et al. 2021 ; Ezugwu et al. 2022 ). These approaches can attempt to learn, e.g., the true non-Euclidean distances between genres played by bands (Mahalanobis 1936 ; Xie et al. 2016 ). The problem is that this learning is severely limited if the space is defined by a complex network (Bronstein et al. 2017 ), as is the case here. Therefore, one would have to use Graph Neural Networks (GNN) (Scarselli et al. 2008 ; Wu et al. 2022 ; Zhou et al. 2020 ). However, GNNs focus on node analysis (Bo et al. 2020 ; Tsitsulin et al. 2020 ; Bianchi et al. 2020 ; Zhou et al. 2020 ), usually via finding the best way of creating node embeddings (Perozzi et al. 2014 ; Hamilton et al. 2017 ). GNNs only use node attributes for the purpose of aiding the analysis of nodes rather than analyzing the attributes themselves (Perozzi et al. 2014 ; Zhang et al. 2019 ; Wang et al. 2019 ; Lin et al. 2021 ; Cheng et al. 2021 ; Yang et al. 2023 ). Previous research shows that, when focusing on node attributes rather than on nodes, the techniques we use here are more suitable than adapting GNNs developed with a different focus (Damstrup et al. 2023 ).
Another class of alternative to deal with this data richness is to use hypergraphs (Bretto 2013 ) and high order networks (Bianconi 2021 ; Benson et al. 2016 ; Lambiotte et al. 2019 ; Xu et al. 2016 ). With these techniques, it is possible to analyze relationships involving multiple actors at the same time—rather than only dyadic relationships like in simpler network representations—and encode path dependencies—e.g. using high order random walks where a larger portion of the network is taken into account to decide which node to visit next (Kaufman and Oppenheim 2020 ; Carletti et al. 2020 ). While a comparative analysis between these techniques and the ones used in this paper is interesting, in this paper we exclusively focus on the usefulness of techniques based on node attribute analysis. We leave the comparison with hypergraphs and high order networks as a future work.
Our analysis shows that the node attribute techniques can help addressing a number of interesting research tasks in cultural data analytics. We show that we are able to describe the eclecticism required by music genres—or expressed in time periods –, by how dispersed they are on the music network. We can determine the geographical connection of specific genres, by estimating their correlation not merely based on how many bands from a specific region play a genre, but how bands not playing that genre relate with those that do. We can create new genre categories by looking at how close they are to each other on the music network. We can apply the same logic to discover eras in Italian music production, clustering years into coherent periods.
Finally, we show that our node attribute analysis rest on some assumptions that are likely to be true in our network—that bands tend to share artists if they play similar genres, in similar time periods, and hailing from similar regions.
We release our data as a public good freely accessible by anyone (Coscia 2024 ), along with all the code necessary to reproduce our analysis. Footnote 1
In this section we present our data model and a summary description of the data’s main features. Supplementary Material Section 1 provides all the details necessary to understand our choices when it comes to data collection, cleaning, and pre-processing.
To obtain a coherent network and to limit the scope of our data collection, we focus exclusively on the record credits from published Italian bands. The data from this project comes from crowd-sourced user-generated data. We mainly use Wikipedia Footnote 2 and Discogs. Footnote 3 We should note that these sources have a bias favoring English-speaking productions. While this bias does not affect our data collection too much, since we focus on Italy without comparing it to a different country/culture, it makes it more likely that there are Italian records without credits, or that are simply missing.
Our bipartite network data model. Artists in blue, bands in red. Edges are labeled with the first-last year in which the collaboration was active. The edge width is proportional to the weight, which is the number of years in which the artist participated to records released by the band
Figure 1 shows our data model, which is a bipartite network \(G = (V_1, V_2, E)\) . The nodes in the first class \(V_1\) are artists. An artist is a disambiguated physical real person. The nodes in the second class \(V_2\) are bands, which are identified by their name. Note that we consider solo artists as bands, and they are logically different from the artist with the same name. Note how in Fig. 1 we have two nodes labeled “Ginevra Di Marco”, one in red for the band and the other in blue for the artist.
Each edge \((v_1, v_2, t)\) —with \(v_1 \in V_1\) and \(v_2 \in V_2\) —connects an artist if they participated in a record of the band. The bipartite network is temporal. Each edge has a single attribute t reporting the year in which this performance happened. This implies that there are multiple edges between the same artist and the same band, one per year in which the connection existed—for notation convenience, we can use \(w_{v_1,v_2}\) to denote this count for an arbitrary node pair \((v_1, v_2)\) , since it is equivalent to the edge’s weight.
We have multiple attributes on the band. The attributes are divided in three classes. First, we have genres. We recover from Discogs 477 different genres/styles that have been used by at least one band in the network. Each of these genres is an attribute of the band, and the value of the attribute is the number of records the band has released with that genre. We use S to indicate the set of all genres, and show an example of these attributes in Table 1 (first section). The second attribute class is the one-hot encoded geographical region of origin, with each region being a binary vector equal to one if the band originates from the region, zero otherwise. We use R to indicate the set of regions. Table 1 (second section) shows a sample of the values of these attributes. The final attribute class is the activity status of a band in a given year—with Y being the set of years. Similarly to the geographical region, this is a one-hot encoded binary attribute. Table 1 (third section) shows a sample of the values of these attributes.
For the remainder of the paper, we limit the scope of the analysis to a projection of our bipartite network. We focus on the band projection of the network, connecting bands if they share artists. We do so to keep the scope contained and show that even by looking at a limited perspective on the data, node attribute analysis can be versatile and open many possibilities. Supplementary Section 2 contains summary statistics about the bipartite network and the other projection—connecting artists with common bands.
There are many ways to perform this projection (Newman 2001 ; Zhou et al. 2007 ; Yildirim and Coscia 2014 ), which result in different edge weights. Here we weight edges by counting the number of years a shared artist has played for either band. Supplementary Material Section 1 contains more details about this weighting scheme. Since we care about the statistical significance—assuming a certain amount of noise in user-generated data—we deploy a network backboning technique to ensure we are not analyzing random fluctuations (Coscia and Neffke 2017 ).
Table 2 shows that the band projection has a low average degree and density, with high clustering coefficient and modularity—which indicate that one can find meaningful communities in the band projection. These are are typical characteristics of a wide variety of complex networks that can be found in the literature.
Table 3 summarizes the top 10 bands according to three standard centrality measures: degree, closeness, and betweenness centrality. Degree is biased by the density of the hip hop cluster—which, as we will see, is a large quasi-clique, including only hip hop bands. Closeness is mostly dominated by alternative rock bands, as they happen to be in the center of mass of the network. The top bands according to betweenness are those bands that are truly the bridges connecting different times, genres, and Italian regions. Note that we analyze the network as a cumulative structure, therefore these centrality rankings are prone to overemphasize bands that are in the central period of the network, as they naturally bridge the whole final structure. In other words, it is harder to be central for very recent or very old bands.
The temporal component of the band projection. Each node is a band. Edges connect bands with significant number of artist overlap. The edge’s color encodes its statistical significance (in increasing significance from bright to dark). The edge’s thickness is proportional to the overlap weight. The node’s size is proportional to its betweenness centrality. The node’s color encodes the average year of the band in the data—from blue (low year, less recent) to red (high year, more recent)
We visualize the band projection to show visually the driving forces behind the edge creation process: temporal and genre assortativity. For this reason we produce two visualizations. First, we take on the temporal component in Fig. 2 . The network has a clear temporal dimension, which we decide to place on a left-to-right axis in the visualization, going from older to more recent.
Second, we show the genre component in Fig. 3 , which instead causes clustering—the tendency of bands playing the same genre to connect to each other more than with any other band. For simplicity, we focus on the big three genres—pop, rock, and electronic—plus hip hop, since the latter creates the strongest and most evident cluster notwithstanding being less popular than the other three genres. For each node, if the band published more than a given threshold records in one of those four genres, we color the node with the most popular genre among them. If none of those genres meets the threshold, we count the band as playing an “other” generic category.
The genre component of the band projection. Same legend as Fig. 2 , except for the node’s color. Here, color encodes the dominant genre among pop (green), rock (red), electronic (purple), hip hop (blue), and other (gray)
This node categorization achieves a modularity score of 0.524, which is remarkably high considering that it uses no network information at all—and it is not a given that this is the correct number of communities. This is a sign that the network is strongly assortative by genre. With our division in four genres plus other, we observe an assortativity coefficient of 0.689, which is quite high. The assortativity coefficient for the average year of activity is even higher (0.91).
We omit showing the network using the regional information on the bands for two reason. First, there are too many regions (20) to visualize them by using different colors for nodes. Second, the structural relationship between the network and the regions is weaker—the assortativity coefficient being 0.223—which would lead to a less clear visualization.
From the figures and the preliminary analysis, it appears quite evident that the structure of the network has a set of complex and interesting interactions with time, genres, and, to a lesser extent, geography. This means that it is meaningful to use the network structure to estimate the relationship between genres, time, and space. This is the main topic of the paper and we now turn our attention to this analysis.
In this section we investigate a number of potential research questions in cultural data analytics. Each of them is tackled with a different node attribute analysis technique: network variance (Devriendt et al. 2022 ), network correlation (Coscia 2021 ; Coscia and Devriendt 2024 ), and Generalized Euclidean distance (Coscia 2020 )—which is at the basis of node attribute clustering (Damstrup et al. 2023 ) and era discovery. Supplementary Material Section 3 explains in details each of these methods.
When focusing on the genre attributes of the nodes, their network variance can tell us how concentrated or dispersed they are in the network. A disperse genre means that the bands playing that genre do not share artists, not even indirectly: they are scattered in the structure. Vice versa, a low-variance genre implies that there is a clique of artist playing it, and they are shared by most of the bands releasing records with that particular genre. Table 4 reports the five most (and least) concentrated genres in the network.
We only focus on genres that have a minimum level of use, in this specific case at least 1% of bands must have released at least one record using that specific genre. The values of network variance should be compared with a null version of the genre—the values themselves do not tell us whether they are significant or if we would get that level of variance simply given the popularity of the genre. For this reason we bootstrap a pseudo p-value for the variance.
Let’s assume that \(\mathcal {S}\) is a \(|V| \times |S|\) genre matrix. The \(\mathcal {S}_{v,s}\) entry tells us how many records with genre s the band v has published. We can create \(\mathcal {S}'\) , a randomized null version of \(\mathcal {S}\) . In \(\mathcal {S}'\) , we ensure that each null genre has the same number of records as it has in \(\mathcal {S}\) . We do so by extracting with replacement at random \(\sum \limits _{v \in V} \mathcal {S}_{v,s}\) bands for genre s . The random extraction is not uniform: each band has a probability of being extracted proportional to \(\sum \limits _{s \in S} \mathcal {S}_{v,s}\) . In this way, \(\mathcal {S}'\) has the same column sum and similar row sum as \(\mathcal {S}\) . In other word, we randomize \(\mathcal {S}\) preserving the popularity of each genre and each band. Then, we can count the number of such random \(\mathcal {S}'\) s in which the null genre has a higher (lower) variance than the observed genre.
Table 4 shows that stoner rock has a high and significant variance, indicating that bands playing stoner rock have a low degree of specialization. This can be contextualized by the fact that stoner rock was tried out unsystematically by a few unrelated bands, ranging from heavy metal to indie rock. On the other hand, many variants of heavy metal have low variance. This can be explained by the fact that heavy metal is a niche genre in Italy, and all bands playing specific heavy metal variants know each other and share members.
Two genres ( a Hip Hop, b Beat) with different variance. Node size, node definition, and edge thickness, color, and definition is the same as Fig. 2 . The color is proportional to the genre-band node attribute value, with bright colors for low values and dark colors for high values
In Fig. 4 we pick two representative genres—Hip Hop and Beat—which both have the same relatively high popularity in number of bands playing them, and have a significant (low or high) variance and we show how they look like on the network. The figure shows that the variance measure does what we intuitively think it should be doing: the Hip Hop bands have low variance and therefore strongly cluster in the network, while the Beat bands are more scattered.
We are not limited to the calculation of variances for genres: we can perform the same operation for the years. If the variance of a genre tells us how diverse the set of bands playing is, the variance of a year can tell us how diverse the year was. Figure 5 shows the evolution of variances per year. We test the statistical significance of the observed variance value by shuffling the values of the node attribute for a given year a number of times, testing whether the observation is significantly higher, lower, or equal to this expectation.
The network variance (y axis) for a given decade (x axis). Background color indicates the statistical significance: red = lower than expected, green = higher than expected, white = not significantly different from expectation
From the figure we can see that there seems to be two phase transitions. In the first regime, we have an infancy phase with low activity and low variance. The first phase transition starts in the year 1960 and brings the network to a second regime of high activity and high variance. After the peak around the year 1980, a second phase transition introduces the third regime from the mid 90 s until the present, with high activity but low variance. In the latter years, we see hip hop cannibalizing all genres and compressing the record releases to its tightly-knit cluster.
We can now shift our attention from describing a single node attribute at a time—its variance as we saw in the previous sections—to describing the relationships between pairs of attributes. In this section, we do so by calculating their network correlation. Specifically, we want to make a geographical analysis. The ultimate aim is to answer the question: what are some particular strong genre-region associations? We can answer the question by calculating the network correlation between two node attributes, one recording the genre intensity for a band and the other a binary value telling us whether the band is from a specific region or not. The network correlation is useful here, because it grows not only if there are a lot of bands playing that specific genre in that specific region, but also if the other bands in the region that do not play that genre are close in the network to—i.e. share members with—bands playing that genre.
In Table 5 we report some significant region-genre associations. For each region, we pick the most popular genre in the network to which they correlate at a significant level—and they have the highest correlation among all other regions that correlate significantly to that genre. The significance is estimated via bootstrapping, by randomly shuffling the region vector—i.e. changing the set of bands associated to the region while respecting its size. Table 5 does not report a genre for all regions, because for some regions there was no genre satisfying the constraints. Note that some regions might correlate more strongly or more significantly with a genre that is not reported in the table, but we omit it if there was another region with a stronger correlation for that genre.
When we measure the pairwise distance between all node attributes systematically we can cluster them hierarchically. Here, we do such a network-based hierarchical clustering on the music genres and styles as recorded by Discogs. The aim is to see whether we can find groups of genres that are similar to each other, potentially informing a data-driven musical classification. Figure 6 shows a bird’s eye view of the hierarchical clustering, with the similarity matrix and the dendrogram.
The hierarchical genre clusters. The heatmap shows the pairwise similarity among the genres—from low (dark) to high (bright) similarity. The dendrograms show the hierarchical organization of the clusters
To make sense of it, we have selected some clusters, for illustrative purposes only. Table 6 shows what genres and styles from Discogs end up in the color-highlighted clusters from Fig. 6 . We can see that the clusters include similar genres which make as a coherent set of more general music styles. The figure also highlights that there is a hierarchical structure of music styles, with meaningful clusters-within-clusters, and clear demarcation lines between groups and subgroups.
Recall that these clusters are driven exclusively by the network’s topology and do not use any feature coming from the songs themselves. This means that using a network of shared members among bands is indeed insightful in figuring our the related genres these bands play. Therefore, network-based clustering has the potential to guide the definition of new musical classifications.
We now look at the eras of Italian music we can discover in the data. Figure 7 shows the dendrogram, connecting years and groups of years at a small network distance to each other. Each era we identify colors its corresponding branch in the dendrogram. We avoid assigning an era for years pre-1906 and post-2018, due to issues with the representativeness of the data. We also notice that the 1938–1945 period is tumultuous, with many small eras in a handful of years, which is understandable given the geopolitical situation at a time, and so we ignore that period as well.
The eras dendrogram. Clusters join at a height proportional to their similarity level (the more right, the less similar). Colors encode the detected eras with labels on the left
To make sense of temporal clustering, the standard approach in the literature would be to compare counts of activities across clusters. However, that would ignore the role of the network structure. In our framework, we can characterize eras applying the same logic used to find them. We calculate the network distance between a node attribute representing the era and each genre. The era’s node attribute simply reports, for each band, a normalized count of records they released within the bounds of that era. We normalize so that each era attribute sums to one, to avoid overpowering the signal with the scale of the largest and most active eras.
Then, for each era, we report the list of genres that have the smallest distance with that era. Note that some genres might still have a small distance with other eras, but we only report the smallest. These are the genres we use to label the eras in Fig. 7 . These genres are not the most dominant in that era—in almost all cases, pop and rock dominate—but they give an intuition of what was the most characteristic genre of the era, distinguishing it from the others.
We can see that the characterization makes intuitive sense, with the classical genres being particularly correlated with the 1906–1916 era. Beat and rock’n’roll are particularly associated to the 1965–1971 period, the dates corresponding to the British Invasion in Italy. Notably, the punk genre has its closest association with the most recent era we label, 2006–2017, proving that—at least in Italy—punk is indeed not dead.
Wrapping up the analysis, one key assumption that underpins the analysis we made so far is that the connections in the band projection follow a few homophily rules. We can have meaningful genre (Sect. Genre clusters ) and temporal (Sect. Temporal Clusters ) clusters using our network distance measures only if bands do tend to connect if they have a genre or temporal similarity. Two bands should be more likely to share members if they play similar genres and if they do it at a similar point in time. More weakly, correlations between genres and geographical regions (Sect. Node Attribute Correlation ) also make sense if bands with similar geographical origins also tend to share members more often than expected.
While proving this assumption would require a paper on its own, we can at least provide some evidence in favor of its reasonableness. We do so by running two linear regressions. In the first regression, we want to explain the likelihood of an edge to exist in the band projection with the genre, temporal, and geographical similarity between bands, or:
In this formula:
\(Y_{u,v}\) is a binary variable, equal to 1 if bands u and v shared at least one member, and zero otherwise;
\(\mathcal {G}_{u,v}\) is the genre similarity, which is the cosine similarity between the vectors recording how many records of a given genre bands u and v have published;
\(\mathcal {R}_{u,v}\) is the region similarity, equal to 1 if the bands originate from the same region, and zero otherwise;
\(\mathcal {T}_{u,v}\) is the temporal similarity, in which we take the logarithm of the number of years in which both bands released a record, plus one to counter the issue when the bands did not share a year;
\(\beta _0\) and \(\epsilon \) are the intercept and the residuals.
Note that \(Y_{u,v}\) contains all links with weight of at least one, even those that are not statistically significant and were dropped from our visualizations and analyses from the previous sections. Moreover, it also has to contain all non-links. However, since the network is sparse, it is not feasible to have all non-links in the regression. Thus, we perform a balanced negative sampling: for each link that exists we sample and include in \(Y_{u,v}\) a link that does not.
For \(\mathcal {G}_{u,v}\) we only consider the most popular 38 genres, since sparsely used genres would make bands more similar than what they would otherwise be.
The first column of Table 7 shows the result of the model. The first thing we can see is that we can explain 28.4% of the variance in the likelihood of a edge to exist. This means that 71.6% of the reasons why two bands share a member is not in our data—be it unrecorded social networks, random chance, impositions from labels, etc.
However, explaining 28.4% of the variance in the edge existence likelihood still provides a valid clue that our homophily assumptions should hold. All similarities we considered play a role in determining the existence of an edge: all of their coefficients are positive and statistically significant. Given that these similarity measures do not share the same units—and not even the same domain –, one cannot compare the coefficients directly. However, we can map their contributions to the \(R^2\) by estimating their relative importance (Feldman 2005 ; Grömping 2007 ), which we do in Fig. 8 . From the figure we can see that it is the temporal similarity the one playing the strongest role, closely followed by genre similarity. Spatial similarity, on the other hand, while still being statistically significant, provides little to no additional explanatory power to the other factors.
The relative importance of each explanatory variable to determine the existence of a link between two bands in the band projection
Once we establish that the existence of the connection is related to genre, temporal, and geographical similarity, we can ask the same question about the strength of the relationship between two bands. We apply the same model as before, changing the target variable:
Here, \(\log (W_{u,v})\) is the logarithm of the edge weight. Note that here we only focus on those edges that have a non-zero weight, i.e. those that exist. This is because we do not want this model to try and predict also edge existence, beside its strength, as we already took care of that problem with the previous model.
Table 7 contains the results in its second column. We can see that, also in this case, all three factors are significant predictors of the edge weights. The number of artists two bands share goes up if the two bands play similar genres, with temporal overlap, and if they originate from the same region. The \(R^2\) is noticeably lower, though, which means that \(\log (W_{u,v})\) is harder to predict than \(Y_{u,v}\) .
Figure 9 shows the same \(R^2\) decomposition we did in Fig. 8 for \(Y_{u,v}\) . All explanatory variables explain less variance than in the previous model. Relative to each other, the temporal overlap is the factor gaining more importance than genre similarity.
The relative importance of each explanatory variable to determine the weight of a link between two bands in the band projection
In this paper we have provided a showcase of the analyses and conclusions one could do in cultural data analytics by using node attribute analysis. We focused on the case study of Italian music from the past 120 years. We built a bipartite network connecting artists to bands and then projected it to analyze a band-band network. We have shown how one could identify genres concentrating in such a network, hinting at clusters of bands playing homogeneous genres, using network variance. We have shown a geographical analysis, calculating the network correlation between the region of origin of bands and the genres they play. We have shown how one could create a new music genre taxonomy by performing node attribute clustering on music genre data. We also proposed a novel way of performing era detection in a network, by finding clusters of similar consecutive years, where years are node attributes.
While we believe our analysis is insightful, there are a number of considerations that need to be made to contextualize our work. We can broadly categorize the limitations in two categories: the one relating to the domain of analysis, and the methodological ones.
When it comes to cultural data analytics, we acknowledge the fact that we are working with user-generated data. There is no guarantee that the data is free from significant mistakes, misleading entries, and incompleteness. Furthermore, our results might not be conclusive. We process data semi-automatically, and the coding process is not complete, meaning we miss a considerable amount of the lesser known artists. This also means that there could be biases in the data collection, induced by our decision on the order in which we explore the structure—which might be focusing too much or too little on specific areas of Italian music. As a specific example, in our project we have ignored another potentially rich source of node attributes: information about the music labels/publishers. This is available on Discogs, and we could envision a label to be represented as a node vector, whose entries are the number of records a specific label published for a specific band. We plan to use this information for future work. The coding process is still ongoing, and we expect to be able to complete the network in the near future.
On the methodlogical side, we point out that what we did is only possible in the presence of rich metadata—dozens if not hundreds of node attributes. Networks with scarce node attribute data would not be amenable to be analyzed with the techniques we propose here. However, in cultural data analytics, there is usually a high richness of metadata. Furthermore, many of the node attribute techniques only make sense if the node attributes are somehow correlated with the network structure. The musical genre clustering or the era detection would not produce meaningful results if the probability of two nodes of connecting was not influenced by their attributes—i.e. if the homophily hypothesis does not hold. In our case, the homophily assumption likely holds, as we show in Sect. Explaining the Network .
When considering some specific analyses we performed other limitations emerge. For instance, our era discovery approach exclusively looks at node activities. However, structural changes in the network’s connections also play a key role in determining discontinuities with the past (Berlingerio et al. 2013 ). We should explore in future work how to integrate our node attribute approach with structural methods. When it comes to the use of network variance, how to properly estimate its confidence intervals without using bootstrapping remains a future work. Therefore, the results we present here should be taken with caution, as it might be that some of the patterns we highlight are not statistically significant.
On a more practical side, our node attribute techniques hinge on specific matrix operations. While these can be efficiently computed on GPU using tensor representations, this might put a limit on the size of the networks analyzed, which have to fit in the GPU’s memory.
All data and code necessary to replicate our results are available at http://www.michelecoscia.com/?page_id=2336 and Coscia ( 2024 ).
http://www.michelecoscia.com/?page_id=2336 .
https://it.wikipedia.org .
https://www.discogs.com/ .
Aggarwal CC (2018) Neural networks and deep learning. Springer 10(978):3
Google Scholar
Aljalbout E, Golkov V, Siddiqui Y, Strobel M, Cremers D (2018) Clustering with deep learning: taxonomy and new methods. arXiv preprint arXiv:1801.07648
Bail CA (2014) The cultural environment: measuring culture with big data. Theory Soc 43:465–482
Article Google Scholar
Baym NK, Ledbetter A (2009) Tunes that bind? predicting friendship strength in a music-based social network. Inf Commun Soc 12(3):408–427
Benson AR, Gleich DF, Leskovec J (2016) Higher-order organization of complex networks. Science 353(6295):163–166
Berlingerio M, Coscia M, Giannotti F, Monreale A, Pedreschi D (2013) Evolving networks: eras and turning points. Intell Data Anal 17(1):27–48
Bianchi FM, Grattarola D, Alippi C (2020) Spectral clustering with graph neural networks for graph pooling. In: International conference on machine learning, pp 874–883. PMLR
Bianconi G (2021) Higher-order networks. Cambridge University Press, Cambridge
Book Google Scholar
Bo D, Wang X, Shi C, Zhu M, Lu E, Cui P (2020) Structural deep clustering network. In: Proceedings of the web conference 2020, pp 1400–1410
Bretto A (2013) Hypergraph theory. An introduction. Mathematical Engineering. Springer, Cham 1
Bronstein MM, Bruna J, LeCun Y, Szlam A, Vandergheynst P (2017) Geometric deep learning: going beyond Euclidean data. IEEE Signal Process Mag 34(4):18–42
Brughmans T (2013) Thinking through networks: a review of formal network methods in archaeology. J Archaeol Method Theory 20:623–662
Candia C, Jara-Figueroa C, Rodriguez-Sickert C, Barabási A-L, Hidalgo CA (2019) The universal decay of collective memory and attention. Nat Hum Behav 3(1):82–91
Carletti T, Battiston F, Cencetti G, Fanelli D (2020) Random walks on hypergraphs. Phys Rev E 101(2):022308
Article MathSciNet Google Scholar
Cheng J, Wang Q, Tao Z, Xie D, Gao Q (2021) Multi-view attribute graph convolution networks for clustering. In: Proceedings of the twenty-ninth international conference on international joint conferences on artificial intelligence, pp 2973–2979
Coscia M Italian XX-XXI Century Music. https://doi.org/10.5281/zenodo.13309793
Coscia M (2020) Generalized euclidean measure to estimate network distances. Proc Int AAAI Conf Web Soc Media 14:119–129
Coscia M (2021) Pearson correlations on complex networks. J Complex Netw 9(6):036
Coscia M, Devriendt K (2024) Pearson correlations on networks: Corrigendum. arXiv preprint arXiv:2402.09489
Coscia M, Neffke FM (2017) Network backboning with noisy data. In: 2017 IEEE 33rd international conference on data engineering (ICDE), pp 425–436 . IEEE
Damstrup ASR, Madsen ST, Coscia M (2023) Revised learning via network-aware embeddings. arXiv preprint arXiv:2309.10408
Devriendt K, Martin-Gutierrez S, Lambiotte R (2022) Variance and covariance of distributions on graphs. SIAM Rev 64(2):343–359
Ezugwu AE, Ikotun AM, Oyelade OO, Abualigah L, Agushaka JO, Eke CI, Akinyelu AA (2022) A comprehensive survey of clustering algorithms: state-of-the-art machine learning applications, taxonomy, challenges, and future research prospects. Eng Appl Artif Intell 110:104743
Feldman BE (2005) Relative importance and value. Available at SSRN 2255827
Gleiser PM, Danon L (2003) Community structure in jazz. Adv Complex Syst 6(04):565–573
Grömping U (2007) Relative importance for linear regression in r: the package relaimpo. J Stat Softw 17:1–27
Hamilton W, Ying Z, Leskovec J (2017) Inductive representation learning on large graphs. Adv Neural Inf Process Syst 30
Hohmann M, Devriendt K, Coscia M (2023) Quantifying ideological polarization on a network using generalized Euclidean distance. Sci Adv 9(9):2044
Hristova S (2016) Images as data: cultural analytics and aby warburg’s mnemosyne. Int J Digital Art History (2)
Karjus A, Solà MC, Ohm T, Ahnert SE, Schich M (2023) Compression ensembles quantify aesthetic complexity and the evolution of visual art. EPJ Data Science 12(1):21
Kaufman T, Oppenheim I (2020) High order random walks: beyond spectral gap. Combinatorica 40:245–281
Lambiotte R, Rosvall M, Scholtes I (2019) From networks to optimal higher-order models of complex systems. Nat Phys 15(4):313–320
Lin Z, Kang Z, Zhang L, Tian L (2021) Multi-view attributed graph clustering. IEEE Trans knowl Data Eng
Mahalanobis P (1936) On the generalized distance in statistics. National Institute of Science of India
Manovich L (2020) Cultural analytics. MIT Press, Cambridge
McAndrew S, Everett M (2015) Music as collective invention: a social network analysis of composers. Cult Sociol 9(1):56–80
Mills BJ, Clark JJ, Peeples MA, Haas WR Jr, Roberts JM Jr, Hill JB, Huntley DL, Borck L, Breiger RL, Clauset A (2013) Transformation of social networks in the late pre-hispanic us southwest. Proc Natl Acad Sci 110(15):5785–5790
Moon S-I, Barnett GA, Lim YS (2010) The structure of international music flows using network analysis. New Media Soc 12(3):379–399
Newman ME (2001) Scientific collaboration networks. ii. Shortest paths, weighted networks, and centrality. Phys Rev E 64(1):016132
Pálovics R, Benczúr AA (2013) Temporal influence over the last. FM social network. In: Proceedings of the 2013 IEEE/ACM international conference on advances in social networks analysis and mining, pp 486–493
Pang G, Shen C, Cao L, Hengel AVD (2021) Deep learning for anomaly detection: a review. ACM Comput Surv (CSUR) 54(2):1–38
Park J, Celma O, Koppenberger M, Cano P, Buldú JM (2007) The social network of contemporary popular musicians. Int J Bifurc Chaos 17(07):2281–2288
Pennacchioli D, Rossetti G, Pappalardo L, Pedreschi D, Giannotti F, Coscia M (2013) The three dimensions of social prominence. In: International conference on social informatics, pp 319–332 . Springer
Perozzi B, Akoglu L, Iglesias Sánchez P, Müller E (2014) Focused clustering and outlier detection in large attributed graphs. In: Proceedings of the 20th ACM SIGKDD international conference on knowledge discovery and data mining, pp 1346–1355
Perozzi B, Al-Rfou R, Skiena S (2014) Deepwalk: online learning of social representations. In: Proceedings of the 20th ACM SIGKDD international conference on knowledge discovery and data mining, pp 701–710
Ronen S, Gonçalves B, Hu KZ, Vespignani A, Pinker S, Hidalgo CA (2014) Links that speak: the global language network and its association with global fame. Proc Natl Acad Sci 111(52):5616–5622
Salah AA, Manovich L, Salah AA, Chow J (2013) Combining cultural analytics and networks analysis: Studying a social network site with user-generated content. J Broadcast Electron Media 57(3):409–426
Scarselli F, Gori M, Tsoi AC, Hagenbuchner M, Monfardini G (2008) The graph neural network model. IEEE Trans Neural Netw 20(1):61–80
Schich M, Hidalgo C, Lehmann S, Park J (2008) The network of subject co-popularity in classical archaeology
Stebbins RA (2004) Music among friends: the social networks of amateur musicians’. Popular music: critical concepts in media and cultural studies 1:227–245
Teitelbaum T, Balenzuela P, Cano P, Buldú JM (2008) Community structures and role detection in music networks. Chaos Interdiscipl J Nonlinear Sci 18(4)
Tsitsulin A, Palowitch J, Perozzi B, Müller E (2020) Graph clustering with graph neural networks. arXiv preprint arXiv:2006.16904
Vlegels J, Lievens J (2017) Music classification, genres, and taste patterns: a ground-up network analysis on the clustering of artist preferences. Poetics 60:76–89
Wang C, Pan S, Hu R, Long G, Jiang J, Zhang C (2019) Attributed graph clustering: a deep attentional embedding approach. arXiv preprint arXiv:1906.06532
Wu L, Cui P, Pei J, Zhao L, Guo X (2022) Graph neural networks: foundation, frontiers and applications. In: Proceedings of the 28th ACM SIGKDD conference on knowledge discovery and data mining, pp 4840–4841
Xie J, Girshick R, Farhadi A (2016) Unsupervised deep embedding for clustering analysis. In: International Conference on Machine Learning, pp. 478–487 . PMLR
Xu J, Wickramarathne TL, Chawla NV (2016) Representing higher-order dependencies in networks. Sci Adv 2(5):1600028
Yang S, Verma S, Cai B, Jiang J, Yu K, Chen F, Yu S (2023) Variational co-embedding learning for attributed network clustering. Knowl-Based Syst 270:110530
Yildirim MA, Coscia M (2014) Using random walks to generate associations between objects. PLoS ONE 9(8):104813
Zhang C, Song D, Huang C, Swami A, Chawla NV (2019) Heterogeneous graph neural network. In: Proceedings of the 25th ACM SIGKDD international conference on knowledge discovery & data mining, pp 793–803
Zhou T, Ren J, Medo M, Zhang Y-C (2007) Bipartite network projection and personal recommendation. Phys Rev E 76(4):046115
Zhou J, Cui G, Hu S, Zhang Z, Yang C, Liu Z, Wang L, Li C, Sun M (2020) Graph neural networks: a review of methods and applications. AI Open 1:57–81
Zhou K, Huang X, Li Y, Zha D, Chen R, Hu X (2020) Towards deeper graph neural networks with differentiable group normalization. Adv Neural Inf Process Syst 33:4917–4928
Download references
The author is thankful to Amy Ruskin for the project’s idea, and to Seth Pate and Clara Vandeweerdt for insightful discussions.
Authors and affiliations.
CS Department, IT University of Copenhagen, Copenhagen, Denmark
Michele Coscia
You can also search for this author in PubMed Google Scholar
M.C. designed and performed all experiments, prepared figures, and wrote and approved the manuscript.
Correspondence to Michele Coscia .
Competing interests.
The authors declare that they have no competing interest.
Publisher's note.
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Supplementary material 1., rights and permissions.
Open Access This article is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License, which permits any non-commercial use, sharing, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if you modified the licensed material. You do not have permission under this licence to share adapted material derived from this article or parts of it. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by-nc-nd/4.0/ .
Reprints and permissions
Coscia, M. Node attribute analysis for cultural data analytics: a case study on Italian XX–XXI century music. Appl Netw Sci 9 , 56 (2024). https://doi.org/10.1007/s41109-024-00669-5
Download citation
Received : 17 May 2024
Accepted : 27 August 2024
Published : 05 September 2024
DOI : https://doi.org/10.1007/s41109-024-00669-5
Anyone you share the following link with will be able to read this content:
Sorry, a shareable link is not currently available for this article.
Provided by the Springer Nature SharedIt content-sharing initiative
You are accessing a machine-readable page. In order to be human-readable, please install an RSS reader.
All articles published by MDPI are made immediately available worldwide under an open access license. No special permission is required to reuse all or part of the article published by MDPI, including figures and tables. For articles published under an open access Creative Common CC BY license, any part of the article may be reused without permission provided that the original article is clearly cited. For more information, please refer to https://www.mdpi.com/openaccess .
Feature papers represent the most advanced research with significant potential for high impact in the field. A Feature Paper should be a substantial original Article that involves several techniques or approaches, provides an outlook for future research directions and describes possible research applications.
Feature papers are submitted upon individual invitation or recommendation by the scientific editors and must receive positive feedback from the reviewers.
Editor’s Choice articles are based on recommendations by the scientific editors of MDPI journals from around the world. Editors select a small number of articles recently published in the journal that they believe will be particularly interesting to readers, or important in the respective research area. The aim is to provide a snapshot of some of the most exciting work published in the various research areas of the journal.
Original Submission Date Received: .
Find support for a specific problem in the support section of our website.
Please let us know what you think of our products and services.
Visit our dedicated information section to learn more about MDPI.
Post-occupancy evaluation of the improved old residential neighborhood satisfaction using principal component analysis: the case of wuxi, china.
1.1. research background, 1.2. post-occupancy evaluation, 2. materials and methods, 2.1. study site, 2.2. sampling, 2.3. data collection, 2.3.1. survey instruments and procedure, 2.3.2. ethical considerations, 2.4. data analysis, 2.5. principal component analysis (pca), 3.1. residents’ socioeconomic characteristics, 3.2. main factors, 4. discussion, 4.1. outdoor recreation, 4.2. transport facilities and small parks, 4.3. public service facilities, 4.4. natural environment condition, 4.5. social and human environment, 4.6. safety and security, 4.7. infrastructure and entrance structures, 4.8. public environment and waste facilities, 4.9. limitation of the study, 5. conclusions, supplementary materials, author contributions, data availability statement, conflicts of interest.
Click here to enlarge figure
S/N | General Information of Respondents Profile | Frequency (No) | Total Responses (No) | Percentages (%) | Cumulative (%) | |
---|---|---|---|---|---|---|
1 | Gender | Male | 195 | 49.2 | 49.2 | |
Female | 201 | 396 | 50.8 | 100.0 | ||
2 | Age group | 18–30 | 85 | 21.5 | 21.5 | |
31–45 | 183 | 46.2 | 67.7 | |||
46–55 | 91 | 23.0 | 90.7 | |||
56–65 | 31 | 7.8 | 98.5 | |||
>65 | 6 | 396 | 1.5 | 100.0 | ||
3 | Educational level | Junior high school or under | 27 | 6.8 | 6.8 | |
Senior high school | 123 | 31.1 | 37.9 | |||
College | 211 | 53.3 | 91.2 | |||
Postgraduate and above | 35 | 396 | 8.8 | 100.0 | ||
4 | Marital status | Single | 62 | 15.7 | 15.7 | |
Married | 303 | 76.5 | 92.2 | |||
Divorced | 27 | 6.8 | 99.0 | |||
Widower | 4 | 396 | 1.8 | 100.0 | ||
5 | Occupation/Nature of Employment | Students | 91 | 23.0 | 23.0 | |
Corporate sector | 201 | 50.8 | 73.7 | |||
Public sector | 28 | 7.1 | 80.8 | |||
Self-employed | 26 | 6.6 | 87.4 | |||
Unemployed | 13 | 3.3 | 90.7 | |||
Pensioner | 37 | 396 | 9.3 | 100.0 | ||
6 | Household registration | Wuxi | 326 | 82.3 | 82.3 | |
Out of town | 70 | 396 | 17.7 | 100.0 | ||
7 | Household income (yuan/month/per) | <2000 | 26 | 6.6 | 6.6 | |
2000–4000 | 82 | 20.7 | 27.3 | |||
4000–6000 | 131 | 33.1 | 60.4 | |||
6000–8000 | 84 | 21.2 | 81.6 | |||
>8000 | 73 | 396 | 18.4 | 100.0 | ||
8 | Duration of Residency | Less than 2 years | 23 | 5.8 | 5.8 | |
2–5 years | 70 | 17.7 | 23.5 | |||
Up to 10 years | 121 | 30.6 | 54.0 | |||
Up to 15 years | 81 | 20.5 | 74.5 | |||
More than 15 years | 101 | 396 | 25.5 | 100.0 | ||
9 | Resident population (per household)/Family Size | 1–2 people | 76 | 19.2 | 19.2 | |
3–4 people | 203 | 51.3 | 70.5 | |||
5–6 people | 92 | 23.2 | 93.7 | |||
≥7 people | 25 | 396 | 6.3 | 100.0 | ||
10 | Nature of Housing | Private house | 315 | 79.5 | 79.5 | |
Rented house | 59 | 14.9 | 94.4 | |||
Public house | 22 | 396 | 5.6 | 100.0 |
Modified Outdoor Spaces | Factors | Factor Loading | Eigen Value | Percentage Variance |
---|---|---|---|---|
1. Outdoor recreation | 7.476 | 12.461 | ||
Creating space for playing by children | 0.769 | |||
Creating space for children’s recreational facilities | 0.739 | |||
Creating space for playing by adults | 0.719 | |||
Creating space for outdoor resting | 0.702 | |||
Provision of outdoor seating | 0.701 | |||
Creating space for fitness facilities | 0.695 | |||
Creating space for strolling | 0.665 | |||
Creating space for chess | 0.652 | |||
Creating space for jogging | 0.647 | |||
2. Transport facilities | 4.921 | 8.202 | ||
Creating space for non-motorized charging facilities | 0.748 | |||
Creating space for motor vehicles | 0.739 | |||
Creating space for parking for non-motorized vehicles | 0.720 | |||
Optimizing Pavements | 0.711 | |||
Creating space for motor vehicle charging facilities | 0.702 | |||
Repair of pavement drainage spaces | 0.691 | |||
Creating space for the non-motorized shed | 0.688 | |||
—Optimizing Traffic Organization in the neighborhood | 0.683 | |||
Laying of asphalt pavement | 0.653 | |||
3. Small park | 4.921 | 8.202 | ||
Replacement of other hardscapes | 0.750 | |||
Provision of Pavilion | 0.735 | |||
Provision of recreational seating | 0.726 | |||
Creating space for softscape | 0.704 | |||
Creating space for a garden path | 0.682 | |||
4. Public service facilities | 4.739 | 7.898 | ||
Public transportation is accessibility | 0.766 | |||
Accessibility to educational facilities | 0.753 | |||
Availability of community centers | 0.739 | |||
Accessibility to commercial facilities | 0.733 | |||
Availability of medical stations | 0.715 | |||
5. Natural environment condition | 4.378 | 7.297 | ||
Social environment (public security, organization) | 0.699 | |||
Ecological environment (ecology, pollution, taboos) | 0.676 | |||
Greening and Landscape Environment | 0.670 | |||
Optimizing planning layout | 0.634 | |||
Quiet neighborhood | 0.629 | |||
6. Social and Human Environment | 4.125 | 6.875 | ||
Neighborhood | 0.714 | |||
Level of public participation | 0.697 | |||
Settlement recognition | 0.687 | |||
Continuity of historical and cultural values | 0.674 | |||
Organization of residential activities | 0.632 | |||
7. Outdoor security | 3.363 | 5.605 | ||
Creating space for fire protection gadget | 0.707 | |||
Clearing fire exit and entrance | 0.696 | |||
Clearing firefighting landing | 0.685 | |||
Widening the road to meet the requirements of the fire access lane | 0.682 | |||
8. Outdoor Lighting | 2.658 | 4.431 | ||
Repairing the unit headlights | 0.700 | |||
Creating space for street lamps | 0.675 | |||
Creating space for courtyard lights | 0.662 | |||
9. Entrance structures | 2.505 | 4.175 | ||
Repairing the main entrance gate | 0.675 | |||
Repairing sub-entrance gate | 0.670 | |||
Creating space gate guard post | 0.631 | |||
10. Infrastructure | 2.311 | 3.852 | ||
Repairing the neighborhood wall | 0.673 | |||
Creating space for a ramp for Physically challenged people | 0.647 | |||
Creating space for drying | 0.632 | |||
11. Public Environment | 1.959 | 3.264 | ||
Environmental health (road, open space cleanliness) Cleanliness | 0.635 | |||
Residential exterior styling and color | 0.628 | |||
Availability of public square space | 0.566 | |||
12. Outdoor Waste facilities | 1.738 | 2.897 | ||
Creating space for garbage bin cleaning site | 0.611 | |||
Creating space for Garbage bins | 0.586 | |||
Creating space for garbage collection and disposal/Garbage collection station | 0.559 | |||
Cumulative Variance (Total) | 79.438% |
Factors | Mean | SD |
---|---|---|
Creating space for playing by children | 3.64 | 1.24 |
Creating space for children’s recreational facilities | 3.64 | 1.24 |
Creating space for playing by adults | 3.62 | 1.23 |
Creating space for outdoor resting | 3.66 | 1.23 |
Provision of outdoor seating | 3.71 | 1.23 |
Creating space for fitness facilities | 3.61 | 1.25 |
Creating space for strolling | 3.69 | 1.22 |
Creating space for chess | 3.56 | 1.24 |
Creating space for jogging | 3.75 | 1.22 |
1. Outdoor recreation | ||
Creating space for non-motorized charging facilities | 3.68 | 1.21 |
Creating space for motor vehicles | 3.66 | 1.25 |
Creating space for parking for non-motorized vehicles | 3.67 | 1.21 |
Optimizing Pavements | 3.71 | 1.22 |
Creating space for motor vehicle charging facilities | 3.58 | 1.24 |
Repair of pavement drainage spaces | 3.68 | 1.20 |
Creating space for the non-motorized shed | 3.52 | 1.31 |
—Optimizing Traffic Organization in the neighborhood | 3.65 | 1.20 |
Laying of asphalt pavement | 3.76 | 1.22 |
2. Transport facilities | ||
Replacement of other hardscapes | 3.73 | 1.20 |
Provision of Pavilion | 3.59 | 1.21 |
Provision of recreational seating | 3.61 | 1.26 |
Creating space for softscape | 3.52 | 1.26 |
Creating space for a garden path | 3.62 | 1.23 |
3. Small park | ||
Public transportation is accessibility | 3.73 | 1.18 |
Accessibility to educational facilities | 3.73 | 1.21 |
Availability of community centers | 3.72 | 1.20 |
Accessibility to commercial facilities | 3.73 | 1.17 |
Availability of medical stations | 3.73 | 1.19 |
4. Public service facilities | ||
Social environment (public security, organization) | 3.64 | 1.22 |
Ecological environment (ecology, pollution, taboos) | 3.67 | 1.18 |
Greening and Landscape Environment | 3.65 | 1.21 |
Optimizing planning layout | 3.65 | 1.20 |
Quiet neighborhood | 3.65 | 1.19 |
5. Natural environment condition | ||
Neighborhood | 3.71 | 1.21 |
Level of public participation | 3.62 | 1.21 |
Settlement recognition | 3.62 | 1.23 |
Continuity of historical and cultural values | 3.58 | 1.22 |
Organization of residential activities | 3.57 | 1.22 |
6. Social and Human Environment | ||
Creating space for fire protection gadget | 3.70 | 1.22 |
Clearing fire exit and entrance | 3.71 | 1.20 |
Clearing firefighting landing | 3.71 | 1.20 |
Widening the road to meet the requirements of the fire access lane | 3.68 | 1.21 |
7. Outdoor security | ||
Repairing the unit headlights | 3.62 | 1.25 |
Creating space for street lamps | 3.72 | 1.23 |
Creating space for courtyard lights | 3.63 | 1.22 |
8. Outdoor Lighting | ||
Repairing the main entrance gate | 3.75 | 1.18 |
Repairing sub-entrance gate | 3.67 | 1.21 |
Creating space gate guard post | 3.69 | 1.20 |
9. Entrance structures | ||
Repairing the neighborhood wall | 3.71 | 1.21 |
Creating space for a ramp for Physically challenged people | 3.70 | 1.23 |
Creating space for drying | 3.71 | 1.25 |
10. Infrastructure | ||
Environmental health (road, open space cleanliness) Cleanliness | 3.70 | 1.24 |
Residential exterior styling and color | 3.67 | 1.23 |
Availability of public square space | 3.62 | 1.26 |
11. Public Environment | ||
Creating space for garbage bin cleaning site | 3.66 | 1.24 |
Creating space for Garbage bins | 3.63 | 1.23 |
Creating space for garbage collection and disposal/Garbage collection station | 3.67 | 1.21 |
12. Outdoor Waste facilities |
Outdoor Security | Transport Facilities | Infrastructure | Public Service Facilities Satisfaction | Outdoor Lighting Satisfaction | Outdoor Waste Facilities Satisfaction | Entrance Structures Satisfaction | Outdoor Recreations Satisfaction | Greenery Satisfaction | Small Park Satisfaction | Natural Environment Condition Satisfaction | Public Environment Satisfaction | Social and Human Environment Satisfaction | |
---|---|---|---|---|---|---|---|---|---|---|---|---|---|
Pearson Correlation | 1 | 0.678 ** | 0.610 ** | 0.623 ** | 0.571 ** | 0.635 ** | 0.642 ** | 0.681 ** | 0.606 ** | 0.619 ** | 0.640 ** | 0.636 ** | 0.657 ** |
Sig. (2-tailed) | 0.000 | 0.000 | 0.000 | 0.000 | 0.000 | 0.000 | 0.000 | 0.000 | 0.000 | 0.000 | 0.000 | 0.000 | |
Sum of Squares and Cross-products | 468.225 | 302.673 | 289.672 | 284.851 | 277.290 | 305.773 | 300.650 | 317.976 | 302.269 | 293.593 | 292.379 | 313.364 | 308.440 |
Covariance | 1.185 | 0.766 | 0.733 | 0.721 | 0.702 | 0.774 | 0.761 | 0.805 | 0.765 | 0.743 | 0.740 | 0.793 | 0.781 |
N | 396 | 396 | 396 | 396 | 396 | 396 | 396 | 396 | 396 | 396 | 396 | 396 | 396 |
The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content. |
Zhao, J.; Abdul Aziz, F.; Cheng, Z.; Ujang, N.; Zhang, H.; Xu, J.; Xiao, Y.; Shi, L. Post-Occupancy Evaluation of the Improved Old Residential Neighborhood Satisfaction Using Principal Component Analysis: The Case of Wuxi, China. ISPRS Int. J. Geo-Inf. 2024 , 13 , 318. https://doi.org/10.3390/ijgi13090318
Zhao J, Abdul Aziz F, Cheng Z, Ujang N, Zhang H, Xu J, Xiao Y, Shi L. Post-Occupancy Evaluation of the Improved Old Residential Neighborhood Satisfaction Using Principal Component Analysis: The Case of Wuxi, China. ISPRS International Journal of Geo-Information . 2024; 13(9):318. https://doi.org/10.3390/ijgi13090318
Zhao, Jing, Faziawati Abdul Aziz, Ziyi Cheng, Norsidah Ujang, Hui Zhang, Jiajun Xu, Yi Xiao, and Lin Shi. 2024. "Post-Occupancy Evaluation of the Improved Old Residential Neighborhood Satisfaction Using Principal Component Analysis: The Case of Wuxi, China" ISPRS International Journal of Geo-Information 13, no. 9: 318. https://doi.org/10.3390/ijgi13090318
Article access statistics, supplementary material.
ZIP-Document (ZIP, 3117 KiB)
Mdpi initiatives, follow mdpi.
Subscribe to receive issue release notifications and newsletters from MDPI journals
Multi-reader, multi-case analysis methods (roc, agreement, and other metrics), description, monthly downloads, pull requests, last published, functions in imrmc (2.1.0).
Journal of Cotton Research volume 7 , Article number: 31 ( 2024 ) Cite this article
Metrics details
Bt technology has played significant role in controlling bollworms and increasing cotton yield in earlier days of its introduction, a subsequent decline in yield became apparent over time. This decline may be attributed to various environmental factors, pest dynamics, or combination of both. Therefore, the present biophysical survey and questionnaire were designed to evaluate the impact of Bt cotton on bollworms management and its effect on reducing spray costs, targeting farmers with varied landholdings and educational backgrounds. Additionally, data on farmers' cultivated varieties and the prevalence of bollworms and sucking insects in their fields were recorded. Subsequently, about eleven thousand cotton samples from farmer fields were tested for Cry1Ac , Cry2Ab and Vip3A genes by strip test.
In this analysis, 83% of the farmers planting approved varieties believe that Bt technology control bollworms, while 17% hold contradictory views. Similarly, among farmers cultivating unapproved varieties, 77% agree on effectiveness of Bt technology against bollworms, while 23% disagree. On the other hand, 67% of farmers planting approved varieties believe that Bt technology does not reduce spray costs, while 33% agree with the effectiveness. Similarly, 78% of farmers cultivating unapproved varieties express doubt regarding its role to reduce spray costs, while 22% are in favour of this notion. Differences in opinions on the effectiveness of Bt cotton in controlling bollworms and reducing spray cost between farmers planting unapproved and approved varieties may stem from several factors. One major cause is the heavy infestation of sucking insects, which is probably due to the narrow genetic variation of the cultivated varieties. Additionally, the widespread cultivation of unapproved varieties (21.67%) is also an important factor to cause different opinions on the effectiveness of Bt cotton.
Based on our findings, we propose that the ineffective control of pests on cotton crop may be attributed to large scale cultivation of unapproved varieties and non-inclusion of double and triple transgene technologies in country’s sowing plan. On the basis of our findings, we suggest cotton breeders, regulatory bodies and legislative bodies to discourage the cultivation of unapproved varieties and impure seed. Moreover, the adoption of double and triple Bt genes in cottons with a broad genetic variation could facilitate the revival of the cotton industry, and presenting a promising way forward.
Cotton ( Gossypium hirsutum L.) is an important fibre crop also known as ‘White Gold’ (Ali et al. 2020 ; Jarwar et al. 2019 ). Pakistan earns major part of foreign exchange from cotton crop which contributes significantly towards economy. Pakistan is the 5 th largest cotton producer and 3 rd larger cotton consuming country in the world. It is an important crop for both agriculture and textile industries, and contributes about 0.6% of GDP and 3.1% of value addition in agriculture sector (Ministry of Finance, Government of Pakistan 2023 ). Over the time, cotton production in Pakistan has declined, due to seed adulteration, ineffective use of fertilizers and pesticides, labour mismanagement, unfavourable weather conditions, and irregular input supplies (Ali et al. 2019 ).
Since the introduction of synthetic insecticides, cotton producers relied heavily on those products to control insect pests. Certain factors i.e., insect resistance, secondary pest outbreaks, and pest resurgences cause an increasing application of synthetic insecticides (Trapero et al. 2016 ). The bollworms ( Heliothis and Helicoverpa spp.) and sucking insects ( Bemisia tabaci , Empoasca spp.) developed resistance to traditional pesticides during the 1990’s (Spielman et al. 2017 ). Afterwards, genetically modified (GM) cotton expressing Bacillus thuringiensis (Bt) toxin was introduced to control lepidopteran pests (Jamil et al., 2021a , b ). Resultantly, bollworms which have developed resistance against insecticides were effectively controlled and pesticide use was significantly reduced (Ahmad et al. 2019 ).
First official approval for general cultivation of Bt cotton in Pakistan was granted in 2010 by National Biosafety Committee within the Pakistan Environmental Protection Agency. However, substantial evidence shows cultivation of Bt cotton at farmers field prior to its official approval (Ahmad et al. 2021 ; Almas et al. 2023 ; Razzaq et al. 2021 ), which are Cry1Ac (first-generation cry gene) based and are primarily resistant to lepidopteran pests. In the earlier days of its introduction, the adoption of Bt technology led to a notable surge in cotton production, increasing from 8.7 million bales in 1999 to 14.61 million bales during 2004–2005, within just five to six years period (Rehman et al. 2019 ). Initially, both approved Bt varieties and unapproved ones showed inconsistent and potentially ineffective transgene expression due to the ineffective regulatory system overseeing commercialization of transgenic variety, releasing of new variety and distribution of seed of approved varieties (Ahmad et al. 2019 ). These loopholes in the system, combined with the challenges farmers face in visually assessing varieties genuineness and seed quality during purchase have contributed to the proliferation of spurious or low-quality seeds (Ali et al. 2019 ; Spielman et al. 2017 ).
Now, the area under Bt cotton cultivation is shrinking and the yield is decreased due to increased insect pest infestations (Arshad et al. 2021 ) owing to field evolved resistance in insects (Jaleel et al. 2020 ; Lei et al. 2021 ). Technologically advanced countries like the USA have addressed this issue of insect resistance development by adopting non-Bt cotton refuge systems and pyramiding multiple toxin genes ( Cry1Ac , Cry2Ab , and Vip3A ). However, in developing countries like China, India, and Pakistan, similar strategies were not effectively implemented, causing the field-evolved resistance in bollworms to proliferate (Jamil et al., 2021a , b ; Karthik et al. 2021 ). Another issue faced by farmers planted Bt cotton is increased infestation of sucking pests due to reduced use of pesticides (Ali et al. 2019 ; Shekhawat and Hasmi 2023 ). Hence, it is believed that interplay of various factors i.e. increased insect pest infestation, field evolved resistance, cultivation of unapproved and substandard seeds and adverse weather conditions result in huge loss of cotton production from 14 million bales in 2004–2005 to 4.91 million bales in 2023 (Ministry of Finance, Government of Pakistan 2023 ).
Keeping in view of the above mentioned facts, a survey was designed to evaluate the impact of Bt technology on cotton production across fifteen core cotton growing districts of Punjab, Pakistan and to understand the multifaceted factors affecting cotton production and to find out the root cause of declining of cotton production. In total 400 farmers possessing various landholding and educational background were surveyed to document their views on Bt. cotton's efficacy against bollworms and spray cost. Additionally, 10,986 cotton samples were tested at farmer’s field through strip tests to assess the purity of cotton varieties with respect to Bt ( Cry1Ac , Cry2Ab and Vip3A) genes.
Present study was conducted at Agricultural Biotechnology Research Institute, Ayub Agricultural Research Institute, Faisalabad 38000, Punjab, Pakistan.
The survey was carried out in core cotton growing area of Pakistan, Punjab province. The Punjab is further divided into 36 administrative units called “districts” that vary significantly in cotton production. Out of 36 districts, fifteen were selected, i.e. Faisalabad, Toba Tek Singh, Sahiwal, Pakpattan, Multan, Lodhran, Khanewal, Vehari, Muzaffargarh, Layyah, D.G. Khan, Rajanpur, Bahawalpur, R.Y. Khan and Bahawalnagar, on the basis of acreage under cotton cultivation as outlined in AMIS.PK ( http://www.amis.pk/Agristatistics/DistrictWise/DistrictWiseData.aspx ). Subsequently, 400 farmer fields were selected from all “Tehsils” (sub-administrative unit) with various landholdings and diverse educational backgrounds particularly in the regions with intensive Bt cotton cultivation. The GPS coordinates of each farmer’s location was recorded using Latitude-Longitude App (Financept) and listed in Table 1 .
A structured questionnaire, comprising of six questions was designed to collect data regarding farmer’s demographic factors, farmers' landholdings and viewpoint about effectiveness of Bt technology in controlling cotton bollworms. The questions were: 1) farmers landholding, classified as small (0–10 acres), medium (11–50 acres) or large (above 50 acres). 2) farmers educational background, stratified into uneducated, below matric, matric, bachelor degree, and masters or above qualifications. 3) the efficacy of Bt cotton in controlling bollworms (Yes, No). 4) the role of Bt technology in reducing the frequency of pesticide sprays and respective pesticide cost to farmers (Yes, No). 5) the variety cultivated by farmers (Table S1). 6) insect infestations of i.e. jassid, whitefly, aphid, thrips, mites, American bollworm (AB) and pink bollworm (PB) (low, medium and high). Infestation levels (low, medium, or high) were based on the economic threshold level (ETL) of each insect species. Infestations below the ETL were classified as "low", those comparable to the ETL as "medium", and those exceeding the ETL as "high". The reference point ETLs for various insect species were as follows: jassid (1 nymph or adult per leaf), whitefly (5 adults per leaf), thrips (8–10 adults per leaf), mites (2 adults per leaf), aphid (20 aphids per leaf), AB (4–5 eggs and larvae per 100 plants), and PB (8% infested bolls) (Ali et al. 2019 ; Razaq et al. 2019 ; Rehman et al. 2019 ).
Molecular analysis was performed through strip test for detection and identification of transgenes at four hundred farmer fields, and a total of 10986 samples were tested. At each field, a minimum of 25 samples were collected and tested, at least 10 samples were tested for each variety. Consequently, depending on the number of varieties cultivated by the farmers, more than 25 samples were tested in some fields. The strip tests were performed using QuickStix combo kits (EnviroLogix), which are equipped with built-in antibody coatings for the detection of Cry1Ac , Cry2Ab , and Vip3A transgenes. The procedure for strip test involved pressing the cap of a disposable eppendorf tube onto two leaves to obtain double leaf disc (weighing approximately 20 mg). Subsequently, the leaf samples were finely grinded with the help of disposable pestle by rubbing against the walls of eppendorf tube after adding 0.5 mL of 1X EB2 extraction buffer. The leaf extract and extraction buffer were homogenized by thorough mixing, ensuring the components were evenly combined for accurate and reliable downstream analysis. Following that, the QuickStix combo strips were dipped in eppendorf tube containing leaf extract with arrow pointing downward. After 10 min incubation, bands were developed on strips through antigen-antibody reaction and strips were analysed for the presence of final bands, and results were recorded (Jamil et al., 2021a , b ).
Frequency analysis of Cry1Ac , Cry2Ab , and Vip3A genes was performed using the “ dplyr ” package to streamline data manipulation and summarization. District-wise opinion of farmers on bollworm management and spray cost reduction were analysed using " tidyverse " functions. The association between Bt. technology adoption, farmers' landholding, and education was studied through a heatmap using the " heatmap.2 " function of " gplots " package in R software. The data regarding varieties cultivated by farmers in each district was analysed using stacked bar-chart illustrated by " ggplot2 " package of R software. Lastly, insect pest infestation data was analysed using " dplyr " and " ggplot2 " packages (Ross et al. 2017 ) and Chi-square (χ 2 ) test was performed to check the associations between different qualitative variables using “ chisq.test ” function in R software.
The purposive sampling technique was used assessing the viewpoint of farmers having diverse landholdings and differential educational backgrounds. Landholdings varied among districts showing distinct distribution of farmers with small, medium and large landholdings (Table 1 ). Notably, the highest proportion of large landholders was found in Sahiwal (43%) followed by Faisalabad (33%), Dera Ghazi Khan (31%) and Rajanpur (29%) district. In terms of medium landholders, district Rahim Yar Khan had the highest (74%), while district Layyah had the lowest (35%) proportions. Among small landholders, district Layyah displayed the highest (50%) while district Sahiwal having the lowest (7%) ratio. Overall, 60% of the farmers have medium, 18% owned small and 22% possessed large landholdings (Table 1 ).
Similarly, variability was observed among farmers on the basis of academic background (Table 2 ). The majority of farmers have completed matric (53%), 22% of farmers were below matric (22%), 12% farmers had bachelor degree, 7% farmers had master degree or above qualifications, and merely 6% farmers were uneducated. The district Sahiwal has highest ratio of uneducated farmers (22%) while highest proportion of farmers with below matric qualification was observed in district Dera Ghazi Khan (39%). Besides Dera Ghazi Khan, all other analysed districts have higher proportion of farmers with matric qualification, specifically, district Toba Tek Singh exhibited highest proportion (100%) followed by Pakpattan (70%). Furthermore, district Bahawalpur and district Layyah exhibited highest proportion of farmers among bachelor degree holders (25%) and master degree or above qualification (30%), respectively (Table 2 ).
The varieties planted at farmer fields were noted and verified based on tags issues by Federal Seed Certification and Registration Department (FSC&RD). The varieties planted at farmer fields were compared with the database of approved varieties from the government to identify the approved or unapproved variety. Overall, unapproved varieties were cultivated extensively covering significant area (21.67%). Moreover among approved varieties the top cultivating were IUB-13 (15.22%), BS-15 (12.61%), FH-142 (8.26%) and FH-Lalazar (8.04%). The lowest cultivated variety was MNH-886 (3.45%). In total of 7.27% area were cultivated with other approved varieties. The top 3 area cultivated unapproved cotton varieties were Bahawalnagar (40.73%), Layyah (38.24%), and Bahawalpur (32.40%). Conversely, unapproved variety was not found in Pakpattan or Toba Tek Singh (Fig. 1 ).
Stacked bar-chart showcasing varietal diversity across fifteen cotton growing districts of cotton belt of Punjab, Pakistan; BWN; Bahawalnagar, BWP; Bahawalpur, DGK; Dera Ghazi Khan, FSD; Faisalabad, KWL; Khanewal, LDN; Lodhran, LYA; Layyah, MTN; Multan, MZG; Muzaffargarh, PKPTN; Pakpattan, RJNPR; Rajanpur, RYK; Rahim Yar Khan, SWL; Sahiwal, TTS; Toba Tek Singh, VHR; Vehari
Analysing region-specific cultivation of varieties, it was observed that IUB-13 was the most cultivated variety in Bahawalnagar (12.09%), Bahawalpur (15.10%), Khanewal (19.18%), Multan (30.41%), Muzaffargarh (21.74%), Faisalabad (24.33%), Rahim Yar Khan (21.10%), and Rajanpur (18.49%). FH-142 was the preferred variety in DG Khan (22.51%) and Layyah (10.66%). FH-Lalazar was most commonly cultivated in Lodhran district (25.64%), while BS-18 dominated in Vehari (21.59%). Additionally, BS-15 was prominently cultivated in Toba Tek Singh (51%), Sahiwal (26.00%), and Pakpattan district (25.60%). Toba Tek Singh and Pakpattan districts have least diversity of cultivated varieties (Fig. 1 ).
To understand the genetic landscape of cultivated varieties with respect to transgenes, strip tests were performed for detection and identification of Cry1Ac , Cry2Ab and Vip3A genes. Across fifteen districts, a total of 10,986 cotton samples were tested. The Cry1Ac gene was presented in varying degrees, with highest occurrence (100%) in district Lodhran, Sahiwal, Pakpattan and Toba Tek Singh. Other districts, such as Khanewal, Bahawalpur, Bahawalnagar, Faisalabad, Layyah, Multan, Rajanpur, Rahim Yar Khan and Vehari also reported more than 80% of Cry1Ac gene in farmer fields. In contrast, Dera Ghazi Khan and Muzaffargarh districts displayed relatively lower percentage of Cry1Ac gene, 69% and 78%, respectively (Table 3 ).
The Cry2Ab gene exhibited a relatively low (9%) percentage throughout the survey area, its frequency ranged from 0% in Pakpattan to 15% in Layyah and Toba Tek Singh districts. The frequency of Cry2Ab gene was no more than 10% in Bahawalnagar, Bahawalpur, Faisalabad, Khanewal, Lodhran, Muzaffargarh, Multan, Pakpattan, Rajanpur, Rahim Yar Khan and Sahiwal districts. Further, the third Bt gene Vip3A, which has broad spectrum resistance against lepidopteron pests, was not found in a single tested sample throughout the survey area. In summary, Cry2Ab gene was found throughout the cotton cultivation regions, except Pakpattan, but the percentage was much lower than Cry1Ac gene (Table 3 ).
The pest counting was performed in the survey area for major cotton pests, i.e., AB, PB, whitefly, aphid, jassid, thrips, and mites. The PB infestation was medium level in more than 50% farmers' fields in most districts except Bahawalnagar, Der Ghazi Khan, Khanewal, Pakpattan and Vehari. Lodhran and Toba Tek Singh recorded 50% of field with low level of PB, whereas Pakpattan and Vehari recorded high level PB invasions at more than 50% fields. In case of AB, Lodhran, Muzaffargarh, Pakpattan and Toba Tek Singh exhibited low AB level at all fields. However, in Bahawalpur and Layyah, 14% and 20% fields experienced medium outbreak of AB, respectively. Notably, in Faisalabad, Layyah and Sahiwal districts, high infestation of AB was observed at 12%, 10% and 7% fields, respectively. On an average, 93% of fields from all survey regions observed low AB outbreak (Table 4 ).
The whitefly remains the predominant insect throughout the survey area with high outbreaks at 68% fields on the average. Five districts including Dera Ghazi Khan, Faisalabad, Muzaffargarh, Rajanpur, and Toba Tek Singh were whitefly hotspot areas, with all survey fields recorded high outbreak. Other districts like Bahawalpur, Khanewal, Multan, Layyah, Rahim Yar Khan, and Sahiwal exhibited diverse infestation patterns. Although aphid is one of the most concerned pest, 77% farmer fields reported low outbreak, particularly in Faisalabad, Sahiwal, Pakpattan, and Toba Tek Singh districts with all observed field recorded as low infestation. On the other hand, 64% fields at Bahawalpur and 44% at Muzaffargarh recorded medium level and 28% fields at district Rahim Yar Khan recorded high level outbreak (Table 4 ).
Apart from whitefly and aphid, jassid was another alarming threat in cotton production, showing high level of invasion at 62% farmer fields. The jassid outbreak in all fields of district Faisalabad, Muzaffargarh, Pakpattan, Rajanpur, Sahiwal, and Toba Tek Singh was at high level. The jassid infestation was also high in more than 70% fields of Bahawalnagar, Dera Ghazi Khan, and Vehari districts. The pest counting of mites revealed low infestation at 62% observed fields. Faisalabad and Pakpattan districts had low infestations in all fields, whereas Dera Ghazi Khan and Muzaffargarh districts recorded medium outbreak at 73% and 86% fields, and 50% fields of Toba Tek Singh recorded as high mites outbreak. The thrips outbreak was high at 60% farmer fields on the average, Bahawalpur, Khanewal, Lodhran, Muzaffargarh, Rajanpur, Toba Tek Singh and Vehari were recorded as high outbreak in 78%, 78%, 75%, 74%, 89%, 100% and 69% fields, respectively while all fileds in Faisalabad showed medium outbreak of thrips (Table 4 ).
The Chi-square (χ 2 ) test was performed to check the association of 17 pairs of factors as detailed in Table 5 . The association of transgene with varieties was non-significant, which means the type of Bt cotton either single or double transgenic cultivars is not showing distinct correlation with the approved or unapproved varieties. Moreover, farmer’s education and landholding have no impact on transgene adoption. Furthermore, association of transgene was also non-significant with thrips which indicates that thrips equally affect non-Bt, single Bt gene, or double Bt gene cotton. However, association of transgene with AB, PB, whitefly, aphid, jassid and mite infestation was significant. This indicates that AB and PB attack vary with transgene and these are interlinked. Similarly, whitefly, aphid, jassid and mites infestation also vary on non-Bt., single and double gene Bt. cotton varieties (Table 5 , S2).
Likewise, association of varieties with AB, aphid and mite was non-significant which reveals that there is no statistical difference of AB, aphid and mite infestation between approved and unapproved varieties. On the contrary, association of varieties with PB, whitefly, jassid and thrips was significant, indicating that infestation of PB, whitefly, jassid and thrips vary among approved and unapproved varieties (Table 5 ).
The farmer’s viewpoint on efficiency of Bt technology in controlling bollworms and reducing spray cost in cotton crop was analysed and it was observed that 83% of farmers cultivating approved varieties, believed in Bt cotton's effectiveness against bollworms, while 17% hold the contrary belief. However, variation exists in different district, i.e. farmers from Bahawalpur, Faisalabad, Rahim Yar Khan, and Sahiwal unanimously agreed (100%) on Bt cotton's effectiveness against bollworms. But 50% farmers in Bahawalnagar, 33% in Toba Tek Singh and some in other districts cultivating unapproved varieties were not convinced. On the other hand, 77% of farmer cultivated unapproved varieties have faith in Bt cotton usefulness for controlling bollworms, whereas 23% expressed disbelief. All farmers cultivating unapproved varieties in Bahawalpur, Faisalabad, Rahim Yar Khan, Sahiwal and Toba Tek Singh districts unanimously believed that Bt cotton is effective against bollworms. In contrary to that, 67% farmers in Multan, 43% in Dera Ghazi Khan, 37% each in Layyah and Muzaffargarh, 33% each in Lodhran and Rajanpur, 31% in Vehari, 23% in Bahawalnagar, and 8% in Khanewal cultivating unapproved varieties are not convinced about this claim. It is evident that farmers cultivating approved varieties express higher confidence in the effectiveness of bollworm control by Bt technology as compared to those cultivating unapproved varieties (Table 6 ).
Similarly, examining the impact of Bt cotton on spray cost reduction revealed a complex scenario. Among farmers planting approved varieties, 33% believed that Bt technology has reduced spray costs, while majority (67%) disagree. Particularly, farmers in districts Bahawalnagar, Dera Ghazi Khan, Faisalabad, Muzaffargarh, Rajanpur, Toba Tek Singh and Vehari unanimously disagreed that Bt technology reduced the spray costs, while all farmers at Bahawalpur, Rahim Yar Khan and Sahiwal have opposite views. Likewise, among farmers cultivating unapproved varieties, 22% express confidence in reducing spray costs by introduction of Bt technology, while 78% hold opposite perspective. In Pakpattan and Bahawalpur 100% of farmers growing unapproved varieties believe in reduction of spray costs, while in Multan, Faisalabad, Khanewal, Layyah, Lodhran, Muzaffargarh, Rajanpur, Sahiwal, Toba Tek Singh and Vehari, all farmers disagreed with this notion. Overall, the analysis highlighted diverse opinions among farmers about the impact of Bt cotton on spray cost reduction (Table 6 ).
In the midst of the changing agricultural technology and the persistent challenges faced by cotton farmers, our study delves into the dynamics surrounding the adoption and effectiveness of Bt cotton technology. With a focus on bollworm management and spray cost reduction, our research navigates through the perceptions and practices of farmers with diverse educational backgrounds and landholdings and revealed main factors affecting cotton farming. We unravel the complexities underlying farmer beliefs, technological advancements, and regulatory frameworks, aiming to chart a course towards sustainable solutions for the revitalization of the cotton crop.
We have approached farmers from all cotton growing districts of the Punjab with diverse backgrounds, i.e. possessing varying landholdings (Table 1 ) and different educational backgrounds (Table 2 ) to increase the reliability of the results (O'Connell et al. 2022 ). The farmers have been inquired about effectiveness of Bt technology against cotton bollworms and its impact on spray cost. Overall, 60% of the farmers have medium landholdings, 22% farmers owned large landholdings and 18% farmers possessed small landholdings (Table 1 ). Likewise, from education perspective, 53% farmers have matric, 22% farmers are below matric, 12% and 7% farmers have bachelor degree and master degree or above qualifications, whereas 6% farmers were uneducated (Table 2 ) representing a mixed population from each strata of education background and landholdings to obtain meaningful information (Swami and Parthasarathy 2020 ).
These farmers' opinion have been bifurcated into two categories based on cultivation of approved and unapproved varieties. The viewpoint of 83% of farmers cultivating approved varieties is that Bt cotton has controlled the bollworms effectively and 17% have opposite opinion. But among those cultivating unapproved varieties, 77% farmers think that bollworms have been controlled after introduction of Bt cotton and 23% farmers have opposite views (Table 6 ). These findings agrees with the study that both approved and unapproved varieties have significant Bt toxin protein level to control bollworms effectively (Spielman et al. 2017 ). Given that AB and PB infestation are dependent on transgenes (Table 5 ) and have an antagonistic relationship (Table S2), and considering that nearly all cultivated varieties (either approved or unapproved) were transgenic (Table S1), the use of these transgenic varieties is likely the primary factor in controlling bollworms (Kashif et al. 2022 ). Moreover, according to a previous study, unapproved varieties are as effective in controlling bollworms as approved varieties, both expressing transgenes at levels lethal to pests (Cheema et al. 2016 ). However, Jamil et al. ( 2021a , b ) have contradictory viewpoint and believe that, unapproved varieties are the leading cause of resistance due to low Bt. toxin level which providing ideal environment for field evolved resistance (Ahmad et al. 2019 ).
In the earlier years of Bt cotton introduction, farmers were largely convinced about its efficiency to control bollworm invasions as reported in different geographies (Gore et al. 2002 ; Kranthi et al. 2005 ) and Pakistan (Arshad et al. 2009 ). However, with the passage of time, without adoption of some levels of refuge plants (plantation of 10% non-Bt crop as refuge) fields have evolved resistance in bollworms (Shahid et al. 2021 ). The situation was further aggravated due to least or no adoption of double ( Cry1Ac and Cry2Ab ) and triple transgene ( Cry1Ac , Cry2Ab and Vip3A ) technologies (Table 3 ). The double and triple transgene cotton have broad-spectrum resistance by different mode of action and corresponding receptor sites in insect gut (Chen et al. 2017 ; Llewellyn et al. 2007 ). Particularly, the Vip3A gene provides broad-spectrum resistance by encoding Bt toxin that disrupts the digestive system upon ingestion, ultimately leading to insect death. Unlike Cry1Ac , Vip3A gene acts through a different mode of action, making it effective against pests that may have developed resistance to Cry1Ac . This diversity in toxin mechanisms helps enhance the overall efficacy of Bt cotton in managing pest populations and reducing crop damage (Chen et al. 2017 ). Some countries swiftly adopted double and triple gene technologies in the cultivation plan, while Pakistan continues to rely solely on the initially introduced single gene ( Cry1Ac ) Bt cotton, which result in the development of resistance in the field (Tabashnik et al. 2013 ; Tabashnik and Carrière 2017 ).
Analysis of farmers' perspective about the efficacy of Bt technology in reducing spray costs has revealed that more than 50% farmers from both categories (planting approved or unapproved varieties) believe that spray cost has not been reduced upon introduction of Bt technology. Specifically, 33% of farmers cultivating approved varieties affirmed that Bt technology effectively reduces spray costs, while 67% hold a contrary viewpoint. Conversely, among farmers planting unapproved varieties, a higher percentage (78%) of farmers have expressed suspicion regarding the effectiveness of Bt cotton in reducing spray costs, with only 22% supporting this notion (Table 6 ). Farmers hold different views on the effectiveness of Bt cotton against bollworms and its impact on spray costs. Majority of farmers claimed that Bt cotton has successfully controlled bollworms, while they also believe that the introduction of Bt cotton has not reduced spray costs. This is attributed to the increased pressure from sucking insect pests such as whitefly, aphid, jassid, thrips, and mites (Table 4 ), which has led to higher spray costs instead of the anticipated reduction. The sucking pest pressure has been increased after introduction of Bt genotypes owing to the low adaptation to local agro ecological conditions (Lu et al. 2022 ) and narrow genetic base (Jamil et al., 2021a , b ). Therefore, these varieties are more vulnerable to sucking pests compared to earlier genetically diverse varieties, thereby necessitats frequent pesticide spray and nullifys the anticipated reduction in spray costs (Arshad et al. 2009 ).
One significant factor influence farmers' believe on Bt technology is large scale cultivation of unapproved varieties (21.67% area). Particularly, in Bahawalnagar, Layyah and Bahawalpur districts (Fig. 1 ). This may be a leading cause in building farmers' perceptions about Bt. cotton's inefficiency to control bollworms and reducing spray costs, reflecting mismanagement rather than inherent flaws in the technology. Because, during the formal varietal approval process, varieties are passed through certain checks, i.e., disease & insect resistance, adaptability to different geographies, response to different climatic factors and genetic diversity from cultivated varieties (Ahmad et al. 2023a , b ; Iftikhar et al. 2019 ). However, if a variety escape through this process and reach farmers field merely on the basis of high yield, it may be susceptible to bollworms and sucking insects (Kranthi and Stone 2020 ). Furthermore, approved varieties may also have mixing of non-Bt seed as reported in one of our previous study (Jamil et al., 2021a , b ), supressing their genetic potential. Perhaps, all the factors explained above, underscores a deficiency on the part of cotton breeders (both public and private sectors) and regulatory bodies (such as FSC&RD), as they have not effectively regulated the supply of unapproved varieties to farmers, lacking proper check and legislative measures (Shahzad et al. 2022 ).
Different opinions among farmers on the effectiveness of Bt cotton may partly be due to cultivation of unapproved varieties. Moreover, least adoption of double and triple transgene technologies and excessive outbreaks of sucking insects particularly whitefly, jassid and thrips exacerbated the situation. To mitigate these challenges, concerted efforts from cotton breeders and regulatory bodies are imperative. Moreover, there is a need to promote and disseminate the latest Bt cotton technologies particularly Cry2Ab and Vip3A genes among farmers on large scale for dissemination of broad-spectrum resistance against bollworms.
All data generated or analysed during this study are included in this published article.
Ahmad S, Cheema HMN, Khan AA, et al. Resistance status of Helicoverpa armigera against Bt cotton in Pakistan. Transgenic Res. 2019;28:199–212. https://doi.org/10.1007/s11248-019-00114-9 .
Article CAS PubMed Google Scholar
Ahmad J, Zulkiffal M, Anwar J, et al. MH-21, a novel high-yielding and rusts resistant bread wheat variety for irrigated areas of Punjab Pakistan. SABRAO J Breed Genet. 2023a;55(3):749–59. https://doi.org/10.54910/sabrao2023.55.3.13 .
Article Google Scholar
Ahmad J, Rehman A, Ahmad N, et al. Dilkash-20: a newly approved wheat variety recommended for Punjab, Pakistan with supreme yielding potential and disease resistance. SABRAO J Breed Genet. 2023b;55(2):298–308. https://doi.org/10.54910/sabrao2023.55.2.3 .
Ahmad S, Shahzad R, Jamil S, et al. Regulatory aspects, risk assessment, and toxicity associated with RNAi and CRISPR methods. In: Abd-Elsalam KA, Lim K, et al., editors. CRISPR and RNAi systems. Cambridge: Elsevier Inc.; 2021. p. 687–721. https://doi.org/10.1016/B978-0-12-821910-2.00013-8 .
Ali M, Kundu S, Alam M, et al. Selection of genotypes and contributing characters to improve seed cotton yield of upland cotton ( Gossypium hirsutum L.). Asian Res J Agri. 2020;13(1):31–41. https://doi.org/10.9734/arja/2020/v13i130095 .
Ali MA, Farooq J, Batool A, et al. Cotton production in Pakistan. In: Jabran K, Chauhan BS, et al., editors. Cotton production. Hoboken: John Wiley & Sons Ltd; 2019. p. 249–76. https://doi.org/10.1002/9781119385523.ch12 .
Almas HI, Azhar MT, Atif RM, et al. Adaptation of genetically modified crops in Pakistan. In: Nawaz MA, Chung G, Tsatsakis AM, et al., editors. GMOs and political stance. Cambridge: Elsevier Inc; 2023. p. 93–114. https://doi.org/10.1016/B978-0-12-823903-2.00002-0 .
Arshad M, Suhail A, Gogi MD, et al. Farmers’ perceptions of insect pests and pest management practices in Bt cotton in the Punjab, Pakistan. Int J Pest Manage. 2009;55(1):1–10. https://doi.org/10.1080/09670870802419628 .
Arshad A, Raza MA, Zhang Y, et al. Impact of climate warming on cotton growth and yields in China and Pakistan: a regional perspective. Agriculture. 2021;11(2):97. https://doi.org/10.3390/agriculture11020097 .
Article CAS Google Scholar
Cheema H, Khan A, Khan M, et al. Assessment of Bt cotton genotypes for the Cry1Ac transgene and its expression. J Agri Sci. 2016;154(1):109–17. https://doi.org/10.1017/S0021859615000325 .
Chen WB, Lu GQ, Cheng HM, et al. Transgenic cotton coexpressing Vip3A and Cry1Ac has a broad insecticidal spectrum against lepidopteran pests. J Invertebr Pathol. 2017;149:59–65. https://doi.org/10.1016/j.jip.2017.08.001 .
Gore J, Leonard B, Church G, et al. Behavior of bollworm (Lepidoptera: Noctuidae) larvae on genetically engineered cotton. J Econ Entomol. 2002;95(4):763–9. https://doi.org/10.1603/0022-0493-95.4.763 .
Ministry of Finance, Government of Pakistan. Agriculture. In: Economic survey of Pakistan. Ministry of Finance, Government of Pakistan. 2023. p. 19–30.
Iftikhar MS, Talha GM, Shahzad R, et al. Early response of cotton ( Gossypium hirsutum L.) genotype against drought stress. Inter J Biosci. 2019;14(2):537–44. https://doi.org/10.12692/ijb/14.2.536-543 .
Jaleel W, Saeed S, Naqqash MN, et al. Effects of temperature on baseline susceptibility and stability of insecticide resistance against Plutella xylostella (Lepidoptera: Plutellidae) in the absence of selection pressure. Saudi J Biol Sci. 2020;27(1):1–5. https://doi.org/10.1016/j.sjbs.2019.03.004 .
Jamil S, Shahzad R, Rahman SU, et al. The level of Cry1Ac endotoxin and its efficacy against H. armigera in Bt cotton at large scale in Pakistan. GM Crops & Food. 2021a;12(1):1–17. https://doi.org/10.1080/21645698.2020.1799644 .
Jamil S, Shahzad R, Iqbal MZ, et al. DNA fingerprinting and genetic diversity assessment of GM cotton genotypes for protection of plant breeders rights. Int J Agric Biol. 2021b;25(4):768–76. https://doi.org/10.17957/IJAB/15.1728 .
Jarwar AH, Wang X, Iqbal MS, et al. Genetic divergence on the basis of principal component, correlation and cluster analysis of yield and quality traits in cotton cultivars. Pak J Bot. 2019;51(3):1143–8. https://doi.org/10.30848/PJB2019-3(38) .
Karthik K, Negi J, Rathinam M, et al. Exploitation of novel Bt ICPs for the management of Helicoverpa armigera (Hübner) in cotton ( Gossypium hirsutum L.): a transgenic approach. Front Microbial. 2021;12:661212. https://doi.org/10.3389/fmicb.2021.661212 .
Kashif N, Cheema HMN, Khan AA, et al. Expression profiling of transgenes ( Cry1Ac and Cry2A ) in cotton genotypes under different genetic backgrounds. J Integr Agric. 2022;21(10):2818–32. https://doi.org/10.1016/j.jia.2022.07.033 .
Kranthi KR, Stone GD. Long-term impacts of Bt cotton in India. Nat Plants. 2020;6(3):188–96. https://doi.org/10.1038/s41477-020-0750-z .
Kranthi K, Dhawad C, Naidu S, et al. Bt-cotton seed as a source of Bacillus thuringiensis insecticidal Cry1Ac toxin for bioassays to detect and monitor bollworm resistance to Bt-cotton. Curr Sci. 2005;88(5):796–800. https://www.jstor.org/stable/24111269 .
CAS Google Scholar
Lei Y, Jaleel W, Shahzad MF, et al. Effect of constant and fluctuating temperature on the circadian foraging rhythm of the red imported fire ant, Solenopsis invicta Buren (Hymenoptera: Formicidae). Saudi J Biol Sci. 2021;28(1):64–72. https://doi.org/10.1016/j.sjbs.2020.08.032 .
Article PubMed Google Scholar
Li H, Wu KM, Yang XR, et al. Trend of occurrence of cotton bollworm and control efficacy of Bt cotton in cotton planting region of southern Xinjiang. Sci Agric Sin. 2006;39(1):199–205. https://www.cabidigitallibrary.org/doi/full/10.5555/20073100228 .
Google Scholar
Llewellyn DJ, Mares CL, Fitt GP. Field performance and seasonal changes in the efficacy against Helicoverpa armigera (Hübner) of transgenic cotton expressing the insecticidal protein vip3A. Agric for Entomol. 2007;9(2):93–101. https://doi.org/10.1111/j.1461-9563.2007.00332.x .
Lu Y, Wyckhuys KA, Yang L, et al. Bt cotton area contraction drives regional pest resurgence, crop loss, and pesticide use. Plant Biotechnol J. 2022;20(2):390–8. https://doi.org/10.1111/pbi.13721 .
O’Connell C, Osmond D. Why soil testing is not enough: a mixed methods study of farmer nutrient management decision-making among US producers. J Environ Manage. 2022;314:115027. https://doi.org/10.1016/j.jenvman.2022.115027 .
Razaq M, Mensah R, Athar HUR. Insect pest management in cotton. In: Jabran K, Chauhan BS, editors. Cotton production. Hoboken: John Wiley & Sons Ltd; 2019. p. 85–107. https://doi.org/10.1002/9781119385523.ch5 .
Razzaq A, Zafar MM, ALI A, et al. Cotton germplasm improvement and progress in Pakistan. J Cotton Res. 2021;4(1):1–14. https://doi.org/10.1186/s42397-020-00077-x .
Rehman A, Jingdong L, Chandio AA, et al. Economic perspectives of cotton crop in Pakistan: a time series analysis (1970–2015)(Part 1). J Saudi Soc Agric Sci. 2019;18(1):49–54. https://doi.org/10.1016/j.jssas.2016.12.005 .
Ross Z, Wickham H, Robinson D. Declutter your R workflow with tidy tools. PeerJ Preprints. 2017;5:e3180v1. https://doi.org/10.7287/peerj.preprints.3180v1 .
Shahid MR, Farooq M, Shakeel M, et al. Need for growing non-Bt cotton refugia to overcome Bt resistance problem in targeted larvae of the cotton bollworms, Helicoverpa armigera and Pectinophora gossypiella . Egypt J Biol Pest Co. 2021;31:1–8. https://doi.org/10.1186/s41938-021-00384-8 .
Shahzad K, Mubeen I, Zhang M, et al. Progress and perspective on cotton breeding in Pakistan. J Cotton Res. 2022;5:29. https://doi.org/10.1186/s42397-022-00137-4 .
Shekhawat SS, Hasmi SK. Safety and benefits of Bt and Bt cotton: factures, refute, and allegations. In: Shekhawat SS, Irsad, Hasmi SK, editors. Genetic engineering. New York: Apple Academic Press; 2023. p. 23–52. https://www.taylorfrancis.com/chapters/edit/10.1201/9781003378273-2 .
Spielman DJ, Zaidi F, Zambrano P, et al. What are farmers really planting? Measuring the presence and effectiveness of Bt cotton in Pakistan. PLoS One. 2017;12(5):e0176592. https://doi.org/10.1371/journal.pone.0176592 .
Article CAS PubMed PubMed Central Google Scholar
Swami D, Parthasarathy D. A multidimensional perspective to farmers’ decision making determines the adaptation of the farming community. J Environ Manage. 2020;264:110487. https://doi.org/10.1016/j.jenvman.2020.110487 .
Tabashnik BE, Carrière Y. Surge in insect resistance to transgenic crops and prospects for sustainability. Nat Biotechnol. 2017;35(10):926–35. https://doi.org/10.1038/nbt.3974 .
Tabashnik BE, Brévault T, Carrière Y, et al. Insect resistance to Bt crops: lessons from the first billion acres. Nat Biotechnol. 2013;31(6):510–21. https://doi.org/10.1038/nbt.2597 .
Trapero C, Wilson IW, Stiller WN, et al. Enhancing integrated pest management in GM cotton systems using host plant resistance. Front Plant Sci. 2016;7:500. https://doi.org/10.3389/fpls.2016.00500 .
Article PubMed PubMed Central Google Scholar
Download references
The authors are thankful to Dr. Shakeel Ahmad, Seed Center, Ministry of Environment, Water and Agriculture, Riyadh, Dr. Muqadas Aleem, Department of Plant Breeding and Genetics, University of Agriculture, Faisalabad, Dr. Waseem Akbar, Maize and Millets Research Institute, Sahiwal for spending significant time on improvement of the technical aspect of our article and Mr. Ahmad Shehzad, Lab Assistant to assist in biophysical survey. Furthermore, Punjab Agriculture Research Board (PARB) for provision of funds for carrying out this study under Grant No. PARB 890.
This work was supported by Punjab Agriculture Research Board, Grant numbers PARB No. 890. Author S.J., S.U.R. and M.Z.I. has received research support from Punjab Agriculture Board.
Authors and affiliations.
Genetically Modified Organisms Development and Testing Laboratory, Agricultural Biotechnology Research Institute, Ayub Agricultural Research Institute, Faisalabad, Punjab, 38000, Pakistan
Shahzad Rahil, Jamil Shakra, Chaudhry Urooj Fatima, Rahman Sajid Ur & Iqbal Muhammad Zaffar
Centre of Excellence for Olive Research and Trainings (CEFORT), Barani, Agricultural Research Institute, Chakwal, Punjab, Pakistan
Iqbal Muhammad Zaffar
You can also search for this author in PubMed Google Scholar
Shahzad R, Jamil S, Rahman SU, and Iqbal MZ Conceived and designed the analysis; Shahzad R, Jamil S, and Chaudhry UF Collected the data; Shahzad R Chaudhry UF and Jamil S Contributed data or analysis tools; Shahzad R and Chaudhry UF Performed the analysis; Shahzad R and Chaudhry UF wrote the paper; Jamil S, Rahman SU, and Iqbal MZ proofread the manuscript. All authors read and approved the final version of the manuscript.
Correspondence to Shahzad Rahil .
Ethics approval and consent to participate.
Not applicable.
Competing interests.
The authors declare that they have no competing interests.
Supplementary table s1. list of varieties cultivated at farmer fields along with their transgene and approval status., 42397_2024_191_moesm2_esm.docx.
Supplementary Table S2. Frequency Table showing the interaction between cotton type (BT and Non-BT) and various pest infestations, including American bollworm (AB), pink bollworm (PB), whitefly, aphid, jassid, and mite.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/ .
Reprints and permissions
Cite this article.
Shahzad, R., Jamil, S., Chaudhry, U.F. et al. In-depth analysis of Bt cotton adoption: farmers' opinions, genetic landscape, and varied perspectives—a case study from Pakistan. J Cotton Res 7 , 31 (2024). https://doi.org/10.1186/s42397-024-00191-0
Download citation
Received : 21 January 2024
Accepted : 18 July 2024
Published : 04 September 2024
DOI : https://doi.org/10.1186/s42397-024-00191-0
Anyone you share the following link with will be able to read this content:
Sorry, a shareable link is not currently available for this article.
Provided by the Springer Nature SharedIt content-sharing initiative
ISSN: 2523-3254
IMAGES
VIDEO
COMMENTS
Data quality issues, including missing or inaccurate data, can hinder analysis. Domain expertise gaps may result in misinterpretation of results. Resource constraints might limit project scope or access to necessary tools and talent. ... Real-world data science case studies play a crucial role in helping companies make informed decisions. By ...
A case study in data science is an in-depth analysis of a real-world problem using data-driven approaches. It involves collecting, cleaning, and analyzing data to extract insights and solve challenges, offering practical insights into how data science techniques can address complex issues across various industries.
Step 1: With Data Analytics Case Studies, Start by Making Assumptions. Hint: Start by making assumptions and thinking out loud. With this question, focus on coming up with a metric to support the hypothesis. If the question is unclear or if you think you need more information, be sure to ask.
Top 25 Data Science Case Studies [2024] In an era where data is the new gold, harnessing its power through data science has led to groundbreaking advancements across industries. From personalized marketing to predictive maintenance, the applications of data science are not only diverse but transformative. This compilation of the top 25 data ...
1. Solving a Data Science case study means analyzing and solving a problem statement intensively. Solving case studies will help you show unique and amazing data science use cases in your ...
Data sources: The research example used is a multiple case study that explored the role of the clinical skills laboratory in preparing students for the real world of practice. Data analysis was conducted using a framework guided by the four stages of analysis outlined by Morse ( 1994 ): comprehending, synthesising, theorising and recontextualising.
A data analytics case study comprises essential elements that structure the analytical journey: Problem Context: A case study begins with a defined problem or question. It provides the context for the data analysis, setting the stage for exploration and investigation.. Data Collection and Sources: It involves gathering relevant data from various sources, ensuring data accuracy, completeness ...
Humana's Automated Data Analysis Case Study. The key thing to note here is that the approach to creating a successful data program varies from industry to industry. Let's start with one to demonstrate the kind of value you can glean from these kinds of success stories. Humana has provided health insurance to Americans for over 50 years.
Data in Action: 7 Data Science Case Studies Worth Reading. The field of data science is rapidly growing and evolving. And in the next decade, new ways of automating data collection processes and deriving insights from data will boost workflow efficiencies like never before. There's no better way to understand the changing nature of data ...
There are 4 modules in this course. This course is the eighth and final course in the Google Data Analytics Certificate. You'll have the opportunity to complete a case study, which will help prepare you for your data analytics job hunt. Case studies are commonly used by employers to assess analytical skills. For your case study, you'll ...
Top 12 Data Science Case Studies. 1. Data Science in Hospitality Industry. In the hospitality sector, data analytics assists hotels in better pricing strategies, customer analysis, brand marketing, tracking market trends, and many more. Airbnb focuses on growth by analyzing customer voice using data science. A famous example in this sector is ...
Defnition: A case study is a research method that involves an in-depth examination and analysis of a particular phenomenon or case, such as an individual, organization, community, event, or situation. It is a qualitative research approach that aims to provide a detailed and comprehensive understanding of the case being studied.
Case study protocol is a formal document capturing the entire set of procedures involved in the collection of empirical material . It extends direction to researchers for gathering evidences, empirical material analysis, and case study reporting . This section includes a step-by-step guide that is used for the execution of the actual study.
The purpose of case study research is twofold: (1) to provide descriptive information and (2) to suggest theoretical relevance. Rich description enables an in-depth or sharpened understanding of the case. It is unique given one characteristic: case studies draw from more than one data source. Case studies are inherently multimodal or mixed ...
Case studies are good for describing, comparing, evaluating and understanding different aspects of a research problem. Table of contents. When to do a case study. Step 1: Select a case. Step 2: Build a theoretical framework. Step 3: Collect your data. Step 4: Describe and analyze the case.
In conclusion, data analytics case studies serve as invaluable tools for businesses seeking growth and innovation. By harnessing the power of data, organizations can make informed decisions ...
5.2 ANALYSIS OF DATA IN FLEXIBLE RESEARCH 5.2.1 Introduction. As case study research is a flexible research method, qualitative data analysis methods are commonly used [176]. The basic objective of the analysis is, as in any other analysis, to derive conclusions from the data, keeping a clear chain of evidence.
In the second data analysis stage we used case study analysis method (Houghton et al., 2014) and identified numerous case examples of the interventions undertaken by intellectual disability nurses ...
For case study analysis, one of the most desirable techniques is to use a pattern-matching logic. Such a logic (Trochim, 1989) compares an empiri-cally based pattern with a predicted one (or with several alternative predic-tions). If the patterns coincide, the results can help a case study to strengthen its internal validity. If the case study ...
question and to create an illustrative data analysis - and the domain expertise needed. As a result, case studies based on realistic challenges, not toy examples, are scarce. To address this, we developed the Open Case Studies (opencasestudies.org) project, which offers a new statistical and data science education case study model.
Data sources The research example used is a multiple case study that explored the role of the clinical skills laboratory in preparing students for the real world of practice. Data analysis was conducted using a framework guided by the four stages of analysis outlined by Morse ( 1994 ): comprehending, synthesising, theorising and recontextualising.
There are 4 modules in this course. This course is the eighth and final course in the Google Data Analytics Certificate. You'll have the opportunity to complete a case study, which will help prepare you for your data analytics job hunt. Case studies are commonly used by employers to assess analytical skills. For your case study, you'll ...
15-17,20,22 Three studies described the cross-case analysis using qualitative data. Two studies reported a combination of qualitative and quantitative data for the cross-case analysis. In each multiple-case study, the individual cases were contrasted to identify the differences and similarities between the cases.
A case study is one of the most commonly used methodologies of social research. This article attempts to look into the various dimensions of a case study research strategy, the different epistemological strands which determine the particular case study type and approach adopted in the field, discusses the factors which can enhance the effectiveness of a case study research, and the debate ...
Cultural data analytics aims to use analytic methods to explore cultural expressions—for instance art, literature, dance, music. The common thing between cultural expressions is that they have multiple qualitatively different facets that interact with each other in non trivial and non learnable ways. To support this observation, we use the Italian music record industry from 1902 to 2024 as a ...
The dynamic interplay of urban expansion, infrastructure developments, and climate fluctuations may introduce changes in flood susceptibility that are not captured in this analysis. Moreover, while the data-driven model employed in this study offers significant advantages over traditional hydraulic models in computational efficiency and ...
When conducting QLR, time is the lens used to inform the overall study design and processes of data collection and analysis. While QLR is an evolving methodology, spanning diverse disciplines (Holland et al., 2006), a key feature is the collection of data on more than one occasion, often described as waves (Neale, 2021).Thus, researchers embarking on designing a new study need to consider ...
The research employed the case study methodology, utilizing structured questionnaires as the primary means of data gathering. The study employed a probability-random sampling strategy to select a diverse and representative group of participants. ... All data analysis was performed using the Statistical Products and Services Solution (SPSS ...
This software does Multi-Reader, Multi-Case (MRMC) analyses of data from imaging studies where clinicians (readers) evaluate patient images (cases). What does this mean? ... Many imaging studies are designed so that every reader reads every case in all modalities, a fully-crossed study. In this case, the data is cross-correlated, and we consider the readers and cases to be cross-correlated ...
Background Bt technology has played significant role in controlling bollworms and increasing cotton yield in earlier days of its introduction, a subsequent decline in yield became apparent over time. This decline may be attributed to various environmental factors, pest dynamics, or combination of both. Therefore, the present biophysical survey and questionnaire were designed to evaluate the ...