Main Information
Overall Assessment:
This bibliographic collection, derived from Scopus, represents a moderately sized dataset (100 documents) spanning from 2007 to 2025. While the annual growth rate is reasonable (8.01%), the relatively low document count necessitates careful interpretation of other metrics. The data suggests a focused, possibly niche, research area, or a collection filtered by specific keywords, institutions, or authors. The fact that the average document age is only 4.01 years suggests the area being researched is emerging, innovative or fast paced.
Key Insights Based on the Statistics:
* Scope and Coverage (Documents & Sources):
* Documents (100): The number of documents is relatively small. This could indicate a specific and focused research area, a very recent area of study, or a highly selective search query. It’s crucial to consider what criteria were used to define this collection. A larger collection would provide more robust statistical power.
* Timespan (2007-2025): The 18-year timespan provides a reasonable window to observe trends, but the document distribution across these years is important. Is the growth consistent, or are there spikes and drops in publication activity?
* Sources (60): The presence of 60 different sources suggests a degree of diversity in the publication outlets relevant to this research area. Analyzing *which* journals and book series these are could reveal the core journals and key publishers in this field.
* Productivity (Authors, Collaboration, Document Types):
* Authors (280): With 280 authors contributing to 100 documents, there’s a good level of engagement within the field. However, it also suggests that a single author does not contribute to many papers.
* Co-Authors per Doc (3.34): The average of 3.34 co-authors per document indicates a strongly collaborative research environment. This is a good sign.
* Single-Authored Docs (4): Only 4 single-authored documents suggest that research in this field is largely collaborative. This is common in many scientific disciplines.
* Document Types: The dominance of “conference papers” (50) and “articles” (35) points to a field where conference presentations are a significant mode of knowledge dissemination, possibly indicating an active and evolving area. The low number of reviews suggests potentially an emerging research area where the synthesis of existing knowledge is still limited. “Books” and “Book Chapters” appear underrepresented, suggesting the research area is not yet mature enough to appear as formal books.
* Impact and Influence (Citations):
* Average Citations per Doc (19.92): This is an important metric. While a direct interpretation requires benchmarking against similar fields and document types, 19.92 citations per document *could* indicate a reasonable level of impact within the specific research area, but it really depends on what field this is. For some fields, this would be excellent; for others, it would be considered low. Crucially, consider citation age. Newer documents will naturally have fewer citations than older ones.
* References (4066): This gives an idea of the depth of the research. Each document, on average, cites around 40 other publications (4066/100=40.66).
* Keywords (ID & DE):
* Keywords Plus (ID): 784 The Keywords Plus (ID) are words or phrases that frequently appear in the cited references of the documents. They are automatically generated by the database (Scopus), and reflect the topics and concepts that are related to the research area.
* Author’s Keywords (DE): 308 The Author’s Keywords (DE) are those chosen by the authors to describe the content of their articles. They reflect the main themes and topics that the authors consider most relevant to their research.
* The Keywords analysis provides insight into the specific themes and concepts being explored. Comparing “Keywords Plus” (ID) with “Author’s Keywords” (DE) can reveal how the field is perceived internally by researchers versus how it’s connected to broader knowledge domains. A large difference can mean the authors of the papers use different keywords than the sources being referenced.
- International Collaboration
* International co-authorships %: 29 29% of international co-authorships suggest a good network of international co-operation
Critical Discussion Points and Further Investigation:
1. Benchmarking: Compare the “Average citations per doc” with the average citation rates for similar document types in *related* fields within Scopus. This provides context for understanding the relative impact.
2. Trend Analysis: Examine how the number of publications and citations have changed over the 2007-2025 period. Is the field growing rapidly? Is the impact (citations) increasing or plateauing?
3. Source Analysis: Identify the most prolific journals and sources in this collection. Are they specialized journals or broader, multidisciplinary ones?
4. Author Network Analysis: Explore the collaboration network among the 280 authors. Are there dominant research groups or institutions?
5. Keyword Analysis: Perform a more detailed analysis of the “Keywords Plus” and “Author’s Keywords.” Identify the most frequent keywords and explore their relationships. Use this to refine your understanding of the research area’s core themes.
6. Document Type Breakdown: Investigate *why* conference papers are so prevalent. Is this a field driven by rapid prototyping and early dissemination of results?
7. Scopus Bias: Acknowledge that Scopus has a particular coverage profile. The results might be different if the same analysis was performed on Web of Science, Dimensions, or Google Scholar.
Guidance for Interpretation:
- Focus on the *relative* values: The absolute numbers (e.g., 100 documents) are less informative than their relationship to each other and to benchmarks in similar fields.
- Consider the limitations of bibliometrics: Citation counts are not a perfect measure of impact. They can be influenced by factors such as field, document type, and time since publication.
- Use the bibliometric data to *guide* your qualitative analysis: The statistics can point you to interesting trends and relationships, but you’ll need to delve into the actual content of the documents to understand the nuances of the research.
By addressing these points and delving deeper into the data, you can create a more nuanced and insightful interpretation of your bibliometric results. Remember to always acknowledge the limitations of the data and to contextualize your findings within the broader research landscape.
Annual Scientific Production
Average Citations Per Year
Three-Field Plot
Overall Structure
The plot is a network visualization, showing connections between authors (‘AU’), cited references (‘CR’), and keywords (‘KW_Merged’) from your Scopus dataset. The thickness of the lines connecting the fields represents the strength or frequency of the association between those elements. A thicker line means that the author, the cited reference, and the keyword appear more often together in the dataset.
Field-Specific Observations:
- AU (Authors – Central Field): The list of authors in this field is pretty standard. These are likely some of the most prolific or impactful authors within your dataset’s scope.
- CR (Cited References – Left Field): Each entry here is a truncated citation string. These represent the sources that the authors in the ‘AU’ field are citing in their publications. Note how the color is slightly different.
- KW\_Merged (Keywords – Right Field): This field displays keywords associated with the publications in your dataset. “Merged” suggests that these might be author-supplied keywords, indexer-assigned keywords (from Scopus), or a combination of both, and that they have been unified in some way.
Interpretation of Connections and Patterns:
1. Key Author Clusters: Examine the authors in the ‘AU’ field and their connections. Authors with many connections to both ‘CR’ and ‘KW\_Merged’ are central figures in the research area covered by your dataset.
2. Influential References: Look at the cited references (‘CR’) that have strong connections to many authors and keywords. These are foundational works in the field. Which papers appear most frequently in the citations? Are there any surprising or unexpected connections?
3. Emerging Themes: Analyze the keywords (‘KW\_Merged’) that are strongly linked to authors and cited references. These represent the key themes and topics being explored in the research area. Are there any emerging trends or shifts in focus?
4. Author-Keyword Relationships: The direct links between authors and keywords can reveal each researcher’s specific area of expertise or research focus. Do any authors consistently associate with specific keywords, indicating their specialization?
5. Citation-Keyword Relationships: The links between cited references and keywords show how certain foundational works relate to specific themes. For example, if a seminal paper on “service innovation” is strongly linked to the keyword “digital technologies,” it suggests that digital technologies are now a significant aspect of service innovation research.
Data-Driven Discussion Points and Questions:
- Interdisciplinarity: Are there connections between keywords from different disciplines, suggesting interdisciplinary research trends?
- Evolution of Research: By examining the publication dates of the cited references and their connections to keywords, you can trace the evolution of research themes over time.
- Missing Links: Are there any authors or keywords that seem disconnected from the main network? This might indicate niche areas or emerging topics that are not yet well-integrated into the broader research landscape.
- Keyword Co-occurrence: The keywords with the strongest connections to each other can reveal the main areas of research and can guide the exploration of the underlying themes.
- What are the main research streams? You could use the graph to classify authors into different research streams, which are defined by the most recurrent keywords, and citations.
Limitations:
- The visualization only shows the *most* frequent connections. Less frequent, but potentially important, relationships may be hidden.
- The interpretation is limited by the quality and coverage of the Scopus database.
- The “merged” nature of the keywords may obscure nuances in the original keyword lists.
To take this analysis to the next level, you could use this plot to inform more detailed analyses. For example:
- Co-citation Analysis: Focus on the ‘CR’ field and analyze the co-citation patterns among the cited references to identify intellectual clusters.
- Keyword Co-occurrence Analysis: Analyze the ‘KW\_Merged’ field to identify the most frequent keyword combinations and explore the underlying themes.
- Author Collaboration Network: Build a separate network visualization focusing on author co-authorship to understand collaboration patterns.
By combining this visualization with additional analyses, you can gain a more comprehensive understanding of the research landscape in your field.
Remember to consult the Biblioshiny documentation and examples for more information on how to interpret and use this type of visualization.
Most Relevant Sources
Core Sources by Bradford’s Law
Sources’ Local Impact
Sources’ Production over Time
Most Relevant Authors
Authors’ Production over Time
Overall Observations:
- Time Span: The plot covers publications roughly from 2018 to 2025. It’s important to note that the data up to 2025 might not be complete, especially for authors with publications in 2025, as it takes time for articles to be indexed and cited.
- Production vs. Impact: The size of the bubbles reflects the number of publications in a given year, while the color intensity reflects the total citations per year (TC/year). This allows us to differentiate between authors who publish frequently versus those whose publications have a higher impact.
- Database: The data comes from SCOPUS, a major bibliographic database. This provides a good starting point for analysis, but it’s important to remember that SCOPUS doesn’t cover every publication.
Individual Author Analysis:
Let’s analyze each author, using the visual information and the provided list of most cited-per-year articles:
1. PARIDA V:
* Timeline: Shows activity from 2020 to 2025, with a clear peak in both publications and citations around 2022. This suggests that their most impactful work occurred around this time.
* Key Publications: “Linking Circular Economy and Digitalisation Technologies…” (2022) stands out with a very high TCpY of 114.5. The other two cited papers, “An Agile Co-Creation Process…” (2020) and “Managing Digital Servitization…” (2022), also show considerable impact.
* Interpretation: Parida V. seems to have made a significant contribution to the intersection of circular economy, digitalization, and servitization, particularly around 2022.
2. KOHTAMÄKI M:
* Timeline: Similar to Parida, shows activity from 2020 to 2025.
* Key Publications: “An Agile Co-Creation Process…” (2020) is particularly highly cited (TCpY 67.2), indicating its significant influence. “Managing Digital Servitization…” (2022) also has a substantial TCpY of 25.5.
* Interpretation: Kohtamäki’s work also appears to be focused on digital servitization and agile co-creation processes, with a strong impact evident from the citation counts. Note the shared publication (“An Agile Co-Creation Process…”) with Parida V and Sjodin D, suggesting collaboration.
3. ZHENG P:
* Timeline: Appears active primarily between 2019 and 2021.
* Key Publications: “A Graph-Based Context-Aware Requirement Elicitation Approach…” (2021) has the highest TCpY (19.4) among their top articles.
* Interpretation: Zheng P’s work seems to be centered around smart product-service systems (PSS) and requirement elicitation. The citation counts suggest a moderate impact within this specific area.
4. CHEN C-H:
* Timeline: Spans from 2021 to 2023.
* Key Publications: “A Graph-Based Context-Aware Requirement Elicitation Approach…” (2021) has a notable TCpY of 19.4. “Artificial Intelligence-Enabled Digital Transformation in Elderly Healthcare Field…” (2023) also shows promising early impact (TCpY 18).
* Interpretation: Chen C-H’s research interests appear to be in smart PSS and AI applications in healthcare, demonstrating a focus on both theoretical and applied aspects.
5. PEZZOTTA G, PIROLA F, and SALA R:
* Timeline: Publications mainly clustered around 2019-2022.
* Key Publications: These authors share the same top three cited articles, all with relatively low TCpY values (around 2.8 and below). This suggests close collaboration on the same research themes.
* Interpretation: Their research seems focused on using NLP for insights discovery in industrial maintenance and decision-support systems for service delivery. The relatively low TCpY values could indicate a niche area or more recent publications that haven’t yet accumulated citations.
6. WANG Z:
* Timeline: Active from 2019 to 2021.
* Key Publications: “A Graph-Based Context-Aware Requirement Elicitation Approach…” (2021) is their most cited paper, with a TCpY of 19.4.
* Interpretation: Wang Z’s research appears to be aligned with Zheng P and Chen C-H, focusing on smart PSS and requirement elicitation, as indicated by the shared highly cited publication.
7. LARSSON T:
* Timeline: Recent activity, primarily in 2024 and 2025.
* Key Publications: All listed publications are from 2024 and 2025, and have low TCpY values.
* Interpretation: Larsson T’s research focuses on AI-driven applications in autonomous vehicles and construction equipment for PSS development. Due to the recent publication dates, it is too early to assess the long-term impact of this work.
8. SJÖDIN D:
* Timeline: Publications from 2020-2025.
* Key Publications: “An Agile Co-Creation Process…” (2020) is their most cited paper (TCpY 67.2), aligning with Kohtamäki and Parida V.
* Interpretation: Sjödin D’s research interests overlap significantly with Kohtamäki and Parida V, focusing on digital servitization and agile co-creation. The 2025 publication suggests ongoing work in AI-enabled PSS.
Further Discussion Points:
- Collaboration: The shared highly-cited papers between certain authors (e.g., Kohtamäki, Parida V, Sjödin D; Zheng P, Chen C-H, Wang Z) indicate strong collaborative relationships and research clusters within this field. This could be explored further to understand the dynamics of these research groups.
- Emerging Trends: The recent publications by Larsson T on AI-driven applications suggest a growing interest in this area. Monitoring the citation impact of these publications in the future would be valuable.
- Limitations: The analysis is based solely on SCOPUS data and TCpY values. It would be beneficial to expand the analysis to include other databases (e.g., Web of Science) and consider other metrics like h-index, co-citation analysis, and bibliographic coupling to gain a more comprehensive view. The age of publications should also be considered when evaluating citation counts. Newer publications will naturally have lower TCpY values.
- Future Research: Based on the analysis, it would be insightful to investigate the specific methodologies and findings of the most highly cited papers to understand the key factors driving their impact. Also, analyzing the research landscape and identifying potential gaps and opportunities for future research would be a useful next step.
By considering these points, you can develop a much more nuanced and insightful interpretation of the bibliometric data, going beyond simple descriptions of publication counts and citation numbers. Remember to always critically evaluate the data and consider its limitations.
Author Productivity through Lotka’s Law
Authors’ Local Impact
Most Relevant Affiliations
Affiliations’ Production over Time
Corresponding Author’s Countries
Overall Observations:
- China leads in publications: China is the most prolific country in this dataset regarding corresponding authors, with 17 articles. However, a significant portion (12) are Single Country Publications (SCP), indicating a strong domestic research focus.
- Varying degrees of international collaboration: The Multiple Country Publication (MCP) ratio varies widely across countries, highlighting different approaches to international collaboration. Some countries, like Finland and Denmark, have a 100% MCP ratio, suggesting a complete reliance on international partnerships for the publications in this dataset. Others, like France, Japan, Austria, Brazil, and others only have single country publication, suggesting that there is no international cooperation in these works.
Detailed Analysis by Country:
- China (17 articles, 29.4% MCP): While being the most productive, China’s MCP ratio is relatively low compared to some other countries. This suggests a robust domestic research ecosystem capable of producing a high volume of publications independently. The 29.4% MCP ratio indicates that there is some engagement in international collaborative research.
- Germany (10 articles, 20% MCP): Germany is the second most productive country. Like China, its MCP ratio is relatively low. This also indicates a strong domestic research base. The 20% MCP suggests a more internally focused research approach within this dataset compared to countries like Sweden, Finland, or Iran.
- Sweden (6 articles, 83.3% MCP): Sweden demonstrates a strong emphasis on international collaboration with a very high MCP ratio. This suggests that Swedish researchers are actively participating in global research networks.
- Finland (4 articles, 100% MCP): Finland exclusively publishes through international collaborations in this dataset. This could suggest reliance on international collaborations, or a strong focus on highly internationalized research areas. The small number of publications should be considered when drawing conclusions.
- Italy (6 articles, 33.3% MCP): Italy’s MCP ratio is moderate, suggesting a balance between domestic and international research efforts.
- Singapore (4 articles, 50% MCP) and United Kingdom (4 articles, 50% MCP): Both countries show an equal distribution between SCP and MCP, indicating a balanced approach to research, with significant engagement in both domestic and international projects.
- Iran (3 articles, 66.7% MCP): Iran demonstrates a high MCP ratio, indicating a strong engagement in international collaborative research.
- France, Japan, Austria, Brazil, EGYPT, KOREA, MEXICO, ROMANIA, SERBIA, SPAIN, and USA (All SCP): The fact that all publications have single country publications, indicates that the publications in the dataset do not present collaborative work.
- Denmark and Hong Kong (All MCP): The fact that all publications present collaborative work indicates a higher emphasis on collaborative projects.
Implications and Discussion Points:
- Research Funding and Policy: The MCP ratio can be influenced by national research funding policies that incentivize international collaborations. The prevalence of SCPs in countries like China and Germany could reflect a strong domestic funding landscape.
- Research Focus and Specialization: A high MCP ratio might indicate that a country specializes in research areas that require international expertise or access to specific resources not available domestically. Consider if the subject area of this bibliometric analysis naturally lends itself to international collaboration.
- Data Source Limitations: The analysis is based on the “corresponding author’s” country. This may not accurately reflect the full extent of collaboration, as other authors from different countries may be involved in SCP publications.
- SCOPUS database limitation: Consider that the collection was downloaded from SCOPUS and does not necessarily reflect the real distribution of international scientific collaboration.
- Impact and Citation Analysis: It would be interesting to examine whether MCPs have a higher citation impact compared to SCPs in this dataset. This could provide insights into the benefits of international collaboration.
- Network Analysis: A network analysis of collaborating countries could reveal specific partnerships and research clusters within the dataset.
Recommendations for Further Research:
- Investigate the specific research areas where these countries are publishing to understand the drivers of collaboration.
- Analyze the citation impact of SCPs vs. MCPs.
- Perform a network analysis to visualize collaboration patterns.
- Compare these findings with similar analyses from other databases (e.g., Web of Science) to assess the robustness of the results.
- Consider the size and scope of the overall dataset when interpreting these results. Small sample sizes can sometimes skew the observed trends.
By considering these factors, you can develop a more nuanced and insightful interpretation of the Corresponding Author’s Country Collaboration Plot and its implications for research within the studied field.
Countries’ Scientific Production
| CHINA | 59 |
| GERMANY | 48 |
| ITALY | 33 |
| SWEDEN | 22 |
| FRANCE | 16 |
| SINGAPORE | 16 |
| UK | 13 |
| AUSTRALIA | 12 |
| FINLAND | 11 |
| USA | 11 |
| SWITZERLAND | 9 |
| BRAZIL | 8 |
| JAPAN | 7 |
| IRAN | 6 |
| ROMANIA | 6 |
| AUSTRIA | 5 |
| NORWAY | 5 |
| SOUTH AFRICA | 5 |
| SOUTH KOREA | 5 |
| EGYPT | 4 |
| INDIA | 4 |
| MEXICO | 4 |
Countries’ Production over Time
Most Cited Countries
Most Global Cited Documents
Most Local Cited Documents
Overall Observations
- Sparse Local Citations: A striking observation is that most articles in this “top locally cited” list have very few (or zero) local citations (LC). This suggests a potential disconnect between the overall research trends (represented by global citations) and the specific focus of the dataset. It also highlights the importance of considering *why* an article might be locally cited or not.
- High Global Citation Variability: There is a wide range of global citations (GC), from 0 to over 400. This indicates that even within this subset of “top local” papers, there’s substantial variation in general scholarly impact.
- Normalization Matters: The NLC (Normalized Local Citations) and NGC (Normalized Global Citations) offer a more nuanced picture. They account for the year of publication. It’s crucial to pay attention to these, as a high GC might simply be because an older paper has had more time to accumulate citations.
- Recent Publications: there is a high number of publications between 2018 and 2024. It will be important to consider this in the interpretation
In-Depth Analysis and Interpretation:
Let’s break down the analysis into a few key areas:
1. Articles with Strong Global Influence (High NGC):
* CHAUHAN C, 2022, TECHNOL FORECAST SOC CHANGE: With an NGC of 11.12 and a GC of 458, this article stands out as having exceptional global impact, even when normalized for its publication year. Although LC is 0, this may be due to the topic of the paper not being in line with the main research focus of the collection. The journal is also not one of the journals included in the collection of the most cited locally
* WANG Z, 2021, INT J PROD RES: This paper has a high global citation count (GC=97) and an NGC of 9.57, indicating a substantial influence in its field. Its LC of 2 suggests some local relevance as well. This one is worth further exploring to see why it’s relevant to both the broader field and the specific focus of this dataset.
* AGRAWAL R, 2022, BUS STRATEGY ENVIRON: This article exhibits a high global citation count (GC=109) and a normalized global citation count of 2.65, indicating significant influence within its field. The local citation count is only LC=1 so it’s not so relevant in the collection
* WALK J, 2023, J CLEAN PROD: This article exhibits a significant global citation count (GC=32) and a normalized global citation count of 2.75, suggesting substantial impact within its domain. However, its local citation count is 0, indicating minimal relevance within the collection’s specific focus.
* LEE C-H, 2023, ADV ENG INF: With GC=54 and NGC=4.64. The article exhibits a significant global influence in its specific field, it doesn’t resonate locally as evidenced by its 0 LC.
* SJÖDIN D, 2020, J BUS RES: The GC = 403 and NGC = 3.66 denote a significant global impact.
2. Articles with Local Relevance (High NLC), even if Globally Lower:
* AEDDULA O, 2024, PROCEDIA CIRP: The NLC is 14 and LC 1. It’s important to note the 2024 publication date – it would be interesting to analyze if the recent articles are all being cited among them but not so much from older publications.
* CHEN D, 2019, PROCEDIA CIRP: With an NLC of 11 and LC of 1, there seems to be local relevance even though globally, it’s not as highly cited (GC=14, NGC=0.89).
* SASSANELLI C, 2022, INT J ENERGY ECON POLICY: This paper stands out with a high NLC of 8 and LC of 2 despite a modest GC of 30 and NGC of 0.73. This suggests that, while not widely cited globally, it’s highly relevant within the specific context of this collection. This would be an interesting paper to examine closely.
* SALA R, 2021, IFIP ADVANCES IN INFORMATION AND COMMUNICATION TECHNOLOGY: This one is interesting because it has a high NLC (7.5) and a low GC (7) and NGC (0.69). This is a good example of an article that is highly relevant to the specific dataset but not as broadly impactful.
3. Articles with Both Local Relevance and Global Influence:
* HEINIS TB, 2018, RES TECHNOL MANAGE: It has the higher LC with 3, and NLC with 4. This indicates its relevance to the research focus of the analyzed dataset.
* WANG Z, 2021, INT J PROD RES: as previously said, it has high values in all the citation parameters considered
4. Articles with Zero Local Citations (LC=0):
* The large number of articles with LC = 0 is a significant point. This could indicate:
* Scope Mismatch: The articles, while relevant to *a* field, aren’t directly aligned with the *specific* research question or themes dominating this particular dataset.
* Citation Practices: Researchers in this specific area might not be citing these particular papers, even if they are broadly relevant.
* Data Limitations: The collection may not fully capture all citations within this specific field.
* Novelty/Emergence: Some newer papers simply haven’t had time to be cited locally yet.
5. Articles with NaN NLC or NGC
* The NaN values for NLC and NGC occur when LC and GC are zero. This means that the document has no impact locally nor globally, or there are issues with the normalization calculation.
Recommendations for Further Analysis:
- Content Analysis: Closely examine the abstracts and keywords of articles with high NLC but low GC to understand *why* they are locally relevant. What specific concepts, methodologies, or findings resonate with this research community?
- Journal Analysis: Identify the journals where the most locally cited articles are published. Is there a core set of journals that define this field?
- Keyword Analysis: Perform keyword co-occurrence analysis to identify the main themes and research areas within the dataset. How do these themes relate to the content of the high-NLC articles?
- Citation Network Analysis: Visualize the citation network within the dataset. Identify the key papers that are being cited by many others.
- Temporal Analysis: Analyze citation trends over time. Are there specific periods when certain articles or themes became more influential within this field?
- Compare to Global Trends: Compare the topics of the most cited local articles to the general trends identified with the most cited global articles. How do they differ, and what does this say about the research focus of your community?
In summary:
This table provides a starting point for a deeper exploration of the research landscape. Understanding the interplay between global and local citations, especially when normalized, can reveal valuable insights into the dynamics of scholarly influence and the specific characteristics of your research field. It highlights the importance of going beyond simple citation counts and considering the context in which research is being cited. The next step is to delve deeper into the content of these articles and the structure of the citation network to uncover more nuanced insights.
Most Local Cited References
Reference Spectroscopy
Overall Interpretation:
The RPYS plot reveals the historical development and key intellectual roots of the research area under investigation, based on data from SCOPUS. The black line illustrates the overall citation activity, showing the number of references cited in the analyzed publications, broken down by year. The red line highlights years where citation frequency significantly deviated from the preceding 5-year median, pinpointing historically important publications that have had a lasting impact on the field.
Key Observations and Interpretations:
- Early Years (Pre-1980s): The plot shows very little citation activity before the 1980s. This could mean the field is relatively new, or that the specific research area of interest emerged more recently.
- 1988: A Foundation Year: The RPYS identifies 1988 as a significant year. The most cited references from that year focus on the “Servitization of Business,” as seen in multiple references from Vandermerwe and Rada. Other influential works from 1988 include “User-Centered System Design” (Anderson, Norman, Draper) and research on Task Load Index (Hart & Staveland) which suggests the development of user-centered design and human factors played a key role. There is also research on electrical stimulation for muscle strengthening, Neural Nets and sport philosophy indicating 1988 was a multidisciplinary year for foundational research.
- 1999: Product-Service Systems Emergence: 1999 shows a clear emphasis on “Product-Service Systems (PSS)“, with multiple citations of Goedkoop’s work. This suggests that PSS began gaining significant traction around this time. There is also research on eco-efficient service innovation (Hockerts) indicating a focus on sustainability in PSS. Other research focuses on bounded rationality and consumer value.
- 2002: Maturation of PSS Concepts: Building on the 1999 foundation, 2002 sees a concentration on “Clarifying the Concept of Product-Service Systems,” primarily driven by Mont’s work. This indicates a period of defining and refining the field. Additional cited works cover topics such as designing product/service systems and the link between PSS and sustainability which suggests these concepts were continuing to develop at this time.
- 2004-2011: The years between 2004 and 2011 show a variety of publications, indicating a more mature and diverse research landscape. Topics include Product Service Systems, Smart Product Service Systems, and Gamification.
- Recent Years (2015-2020): These later peak years include references to Digitalized Product-Service Systems, The Value of Big Data in Servitization, A literature review and a research agenda, and Smart Product-Service Systems, marking the emergence of “smart” and digitally enabled PSS as a prominent area of research. The peak citation activity in recent years suggests a continued interest and active research in this area.
Critical Discussion Points and Further Research:
- Scope of SCOPUS: It’s important to acknowledge that the analysis is limited to SCOPUS-indexed publications. Consider the potential impact of including other databases (Web of Science, Google Scholar) to get a more comprehensive picture.
- Citation Bias: Be mindful of potential citation biases. Highly cited works aren’t always “the best,” but often the most visible or those that have been built upon by many others.
- Thematic Connections: Investigate the connections between the identified peak years. How did the early work in servitization and user-centered design influence the development of PSS, and subsequently, smart PSS?
- Geographic Influence: Examine the authors and institutions associated with these key publications. Were there specific geographic regions or research groups that were particularly influential?
- Methodological Trends: Analyze the research methodologies used in the cited works. Have there been shifts in the dominant research approaches over time (e.g., from conceptual frameworks to empirical studies)?
- Future Directions: Based on the trends identified, what are the likely future directions for research in this area? Are there emerging sub-fields or areas that deserve further investigation? Consider how topics such as AI, circular economy, and servitization models will continue to evolve.
By critically examining the publications associated with these peak years, you can gain a deeper understanding of the historical trajectory, intellectual foundations, and emerging trends within your research area.
Most Frequent Words
WordCloud
TreeMap
Words’ Frequency over Time
Trend Topics
Overall Observations:
- Recent Emergence: Most of the top keywords show significant activity only in the last few years (2018-2024). This suggests that the research area is relatively new or has experienced a surge in interest recently.
- Industry 4.0 and AI Dominance: Terms related to “Industry 4.0” and “Artificial Intelligence” are prominent, indicating a strong focus on these technologies within the analyzed corpus.
- Product-Service Systems Focus: Several keywords revolve around “Product-Service Systems” (PSS), suggesting a significant research interest in this integrated approach to product and service delivery.
- Varied Time Spans: Some terms, like “industrial product service systems”, start earlier (around 2014-2016), while others appear very recently.
Detailed Interpretation by Keyword:
- “Product-Service System” and “Product-Service Systems”: The presence of both singular and plural forms suggests that the concept of Product-Service Systems is a central theme. The recent prominence indicates a growing interest in integrated product and service offerings. Note that there are two “Product service system” topics, one that begin much earlier.
- “Smart Products”: The increasing prominence of “Smart Products” aligns with the broader trend of IoT and connected devices. It indicates research focusing on the design, development, and application of intelligent and interconnected products.
- “Machine Learning”: The rise of “Machine Learning” underscores its increasing application in various domains, including product design, manufacturing, and service delivery. It reflects the integration of AI techniques for automation, optimization, and personalization.
- “Product Design”: The inclusion of “Product Design” as a trend topic is interesting. It can indicate focus on research aimed at improving the processes of product design, taking into account modern technologies and approaches.
- “Industry 4.0”: The presence of “Industry 4.0” as a key trend reflects the increasing adoption of advanced technologies in manufacturing and industrial processes. Research likely focuses on the implementation and impact of Industry 4.0 principles.
- “Artificial Intelligence”: “Artificial Intelligence” is a central topic, and this is supported by its annual frequency.
- “Life Cycle”: The inclusion of “Life Cycle” might suggest research focusing on the entire life cycle of products or services, including design, manufacturing, usage, and end-of-life considerations.
- “Decision Making” and “Decision Support Systems”: The presence of these keywords reflects research on data-driven decision-making processes. It suggests an interest in tools and techniques that support informed decision-making in various contexts.
- “Embedded Systems”: “Embedded Systems” being a trend topic suggests research related to computing systems, where hardware and software are tightly integrated to perform dedicated functions in electronic devices and other machines.
- “Industrial Product Service Systems”: This term, starting earlier than the others, might indicate the initial research focus in this area, with a shift towards more general “Product-Service Systems” and related technologies in recent years.
Critical Discussion Points and Further Investigation:
- Database Bias: The analysis is based on SCOPUS data. Consider whether this database adequately represents all relevant publications in the field. Scopus has known biases (for example, it covers publications in English language better than others), so results may differ if Web of Science or Google Scholar are considered instead.
- Keyword Merging: You’ve used the ‘KW_Merged’ field. Understand how the merging was done. Were synonyms combined? This affects the interpretation.
- Granularity: Using N=3 keywords per year provides a broad overview. Changing this parameter might reveal more nuanced trends or specific sub-topics within each year.
- Contextual Factors: Investigate external factors that might have influenced these trends. For example, government initiatives, funding programs, or technological breakthroughs.
- Future Directions: Based on these trends, what are the potential future research directions? Are there any emerging topics that are not yet represented in the top keywords?
- Compare trends with other analyses: Compare the trend analysis with other bibliometric analysis to validate your findings. For example, compare the results with a co-citation network, or a co-occurrence network.
In summary, this trend topics plot reveals a dynamic research landscape focused on the convergence of Industry 4.0, AI, and Product-Service Systems. The analysis highlights the importance of smart, connected products and the integration of advanced technologies for decision-making and optimization.
Remember to delve deeper into the specific publications associated with these keywords to gain a more comprehensive understanding of the research being conducted in this area.
Clustering by Coupling
Co-occurrence Network
Overall Structure:
The network shows a relatively dense structure, especially around the central term “product-service systems.” This suggests that the literature strongly focuses on this concept, which acts as a hub connecting various related areas. We can see a distribution of terms across the graph, suggesting distinct but interconnected research sub-themes.
Communities (Topics) identified by walktrap:
The Walktrap algorithm detected several communities, each represented by a different color. Here’s a potential interpretation of each:
- Red (Central Cluster): This cluster is dominated by “product-service systems,” “industry 4.0,” “servitization”, “customer satisfaction” “business models”, “product service systems (pss)”. This core cluster represents the central theme of Product-Service Systems and its related aspects in business and industry. This community is heavily focused on the application, management, and optimization of product-service systems, particularly within modern industrial contexts.
- Green (Left Cluster): This cluster includes terms like “semantics,” “smart products,” “digital technologies,” “digitalization”, “cloud computing”, “product design”, “smart product-service system”, and “sustainable development”. This cluster highlights a connection to enabling technologies (cloud computing, digital technologies) and a consideration of design and sustainability aspects within the product-service system context.
- Purple (Top-Middle Cluster): The terms present here are “engineering education,” “sentiment analysis,” “learning systems,” “data analytics,” “machine learning,” “product service system”, “sales”, and “big data.” This community points to the use of advanced analytical techniques (machine learning, data analytics, sentiment analysis) for understanding and improving product-service systems, with potential applications in education and sales.
- Blue (Top-Right Cluster): This smaller cluster includes “health care”, “article”, and “human”. This suggests a branch of research exploring the application of product-service systems within the healthcare sector.
Most Connected Terms and their Relevance:
- “Product-Service Systems”: As the largest node, it’s clearly the central concept. Its high connectivity indicates its importance as a unifying theme.
- “Industry 4.0”: This term is strongly connected to “Product-Service Systems,” indicating a significant trend in the literature focusing on the application of PSS within the context of the Fourth Industrial Revolution. It suggests research on how PSS can be enabled and enhanced by technologies like IoT, AI, and big data.
- “Servitization”: A natural partner to PSS, this term highlights the shift from selling products to selling services and integrated solutions. Its presence suggests that the literature explores the strategies, business models, and organizational changes associated with servitization.
- “Product Design”: The relevance of this term, particularly within the green community, suggests research focuses on integrating service considerations into the product design process to develop more effective and customer-centric PSS offerings.
- “Circular Economy”: This term relates to sustainability and the lifecycle of products, therefore, it means some studies focuses on how PSS can contribute to the goals of a circular economy by designing products for longer lifespans, reuse, and recycling.
Interpretation and Critical Discussion Points:
1. Dominance of the Central Theme: The strong emphasis on “product-service systems” suggests the field is well-established but still evolving. Investigate if the research focuses on theoretical development, case studies, or specific applications.
2. Technology-Driven Innovation: The prominence of “Industry 4.0,” “digital technologies,” and “machine learning” suggests a strong trend toward technology-enabled PSS. Explore how these technologies are being leveraged to create new PSS offerings, improve service delivery, and enhance customer experiences.
3. Sustainability Considerations: The connection between “product design” and “sustainable development” suggests an increasing awareness of the environmental and social impacts of PSS. Examine the literature for studies that address the sustainability challenges and opportunities associated with PSS, such as resource efficiency, waste reduction, and circular economy principles.
4. Industry-Specific Applications: The “health care” cluster hints at specific applications of PSS in different industries. Further investigation might reveal unique challenges and opportunities for PSS implementation in healthcare.
5. Missing Connections: Consider terms that *aren’t* strongly connected. Are there potentially important concepts that are under-represented in the literature? For instance, are ethical considerations (e.g., data privacy, algorithmic bias) adequately addressed in the context of technology-driven PSS?
Further Analysis:
- Temporal Analysis: Consider analyzing the network over different time periods to identify emerging trends and shifts in research focus.
- Author/Country Analysis: Explore the contributions of different authors and countries to the various research themes identified in the network.
- Content Analysis: Supplement the bibliometric analysis with a qualitative content analysis of key publications to gain a deeper understanding of the research findings and methodologies.
By considering these points, you can move beyond a descriptive overview of the network and develop a more nuanced and critical interpretation of the research landscape surrounding product-service systems. Remember to justify your interpretations with evidence from the literature. Good luck!
Thematic Map
Overall Structure and Interpretation
The strategic diagram is a two-dimensional representation of the research landscape, where:
- X-axis (Relevance/Centrality): Indicates the importance or influence of a research area within the overall network. Themes further to the right are more central to the field.
- Y-axis (Development/Density): Reflects the level of development and interconnectedness within a research area. Themes higher up are more well-developed and densely connected.
Based on these axes, the diagram is divided into four quadrants:
- Motor Themes (Top Right): Highly developed and central themes, driving the field forward. These are your core areas of research with a lot of activity and strong connections.
- Basic Themes (Bottom Right): Central to the field but less developed. These are foundational areas that, while important, may not be experiencing rapid growth or innovation currently.
- Niche Themes (Top Left): Well-developed but peripheral themes. These areas are mature but might not be strongly connected to the central research topics. They could represent specialized sub-fields or areas with limited overall impact.
- Emerging or Declining Themes (Bottom Left): Areas that are neither central nor well-developed. These might be new research areas with potential for growth, or established areas that are losing traction.
Cluster Analysis and Interpretation
Now, let’s analyze the clusters based on their position in the strategic diagram and the listed articles. Remember that this analysis is based on keyword co-occurrence, so the “KW_Merged” field is critical.
Motor Themes (Top Right Quadrant):
* Product Design/Learning Systems/Artificial Intelligence/Servitization (Overlapping Clusters – likely a single, large cluster): This is the most important area in this map. This overlapping cluster signals a strong convergence of these areas. The top articles support this interpretation:
* “Product design” is led by high-pagerank articles like DE MOURA LEITE AFS, 2024 and SALA R, 2022, implying a strong focus on this area.
* “Learning Systems” has top articles like TAN K, 2019 and CHEN D, 2019, both from PROCEDIA CIRP, suggesting a strong link to manufacturing and engineering contexts.
* The inclusion of “artificial intelligence” and “product-service system” indicates that the motor themes are focused on the intersection of these topics.
*Interpretation:* The core of this research field seems to be around integrating AI and machine learning into product design, development, and the creation of product-service systems. Servitization, the move towards offering services in conjunction with products, is also a central theme. The fact that the highest ranked AI articles are from PROCEDIA CIRP and SENSORS indicates a strong tie to engineering and practical applications. This quadrant is the driving force in this research landscape.
Basic Themes (Bottom Right Quadrant):
* Circular Economy/Big Data/Business Models: This cluster suggests a focus on the practical application and business implications of circular economy principles, possibly leveraging big data analytics for decision-making and new business model development. This is a core topic, but it could be more developed in the context of this research field.
* The top articles (“CHAUHAN C, 2022,” “HAN Y, 2023,” “DE JESUS PACHECO DA, 2022”) are recent, suggesting current activity and relevance.
*Interpretation:* This is a foundational area that needs further exploration and integration with other themes. The presence of “big data” hints at opportunities for data-driven approaches to circular economy and new business models.
Niche Themes (Top Left Quadrant):
- Industrial Product Service Systems/Industry Adaptability: While related to the motor themes, this cluster is more specialized and less central. It suggests research focused on how industries are adapting to incorporate product-service systems.
*Interpretation:* This cluster represents a more specialized application of product-service systems principles. Its position suggests that while well-developed, it may not be as strongly connected to the central themes driving the overall research field.
- Healthcare: The presence of “healthcare” indicates an application domain for these technologies.
*Interpretation:* The health care cluster shows potential but may require more integration with other themes to move towards the central area.
- Digital Storage/Data Transfer/Data Integration: This cluster likely focuses on the technological aspects of data management, relevant to many of the other themes. The high pagerank of KHORASANI M, 2022 from RAPID PROTOTYPING J indicates a strong interest in the application of these data management principles to design and manufacturing.
*Interpretation:* The strategic map suggests that research on digital storage, data transfer, and integration is relatively developed but not central to the field. To increase its relevance, researchers could focus on connecting these themes to the motor themes (e.g., using advanced data analytics to improve the design and operation of product-service systems).
Emerging or Declining Themes (Bottom Left Quadrant):
- Artificial Intelligence (AI)/Digital Servitization/Smart Building: While “artificial intelligence” appears in the motor themes as well, its presence here suggests a more nascent or less integrated focus on AI *specifically* in relation to smart buildings and digital servitization. The low pagerank of SJÖDIN D, 2020 indicates that this area could be emerging or less researched within the context of the overall keyword network.
*Interpretation:* These areas are potentially emerging. More research is needed to establish their centrality and development within the broader research landscape. For example, researchers could explore how AI can be used to optimize digital servitization strategies in smart buildings, creating stronger links to the motor themes.
Decision Trees and Deep Learning are themes slightly above the horizontal axis in the bottom left quadrant.
*Interpretation:* These areas are potentially becoming less relevant over time, as other more modern methods rise to prominence.
Methodological Considerations & Further Research
- Keyword Choice: The “KW_Merged” field is crucial. Understand exactly how these keywords were merged and if there might be unintended consequences (e.g., over-grouping or loss of nuance).
- Parameters: The parameters used to generate the graph are important. For example, using stemming might change the results, and the `minfreq` parameter affects which keywords are included.
- Database: SCOPUS is a good starting point, but consider comparing results with other databases (Web of Science, etc.) to assess robustness.
- Temporal Trends: This is a snapshot in time. Analyzing strategic maps over different time periods would reveal how themes are evolving.
In summary, this strategic map indicates a strong focus on the integration of AI, machine learning, product design, and servitization, particularly within an engineering and manufacturing context. The “Circular Economy” theme represents a foundational element, and AI’s application to smart buildings and digital servitization are potentially emerging areas. The other themes are relevant to a lesser extent within the context of the dataset. By considering these relationships, you can strategically position your research to contribute to the most impactful areas of the field.
Factorial Analysis
Overall Structure and Dimensions:
* Axes: The map is defined by two dimensions, Dim 1 (20.7% variance explained) and Dim 2 (19.85% variance explained). This indicates that these two dimensions together capture roughly 40% of the total variance in the keyword co-occurrence patterns. While this isn’t extremely high, it’s a reasonable starting point for identifying major themes. Note that MCA often results in lower variance explained per dimension compared to PCA.
* Interpretation of Dimensions:
* Dim 1 (Horizontal): This dimension seems to differentiate between keywords related to “traditional” service/manufacturing approaches and a more “advanced” or “data-driven” approach. Moving from left to right, we see a shift from terms like ‘service-delivery process’, ‘customer satisfaction’, and ‘psmanufacturing companies’ to terms such as ‘semantics’, ‘systems engineering’, ‘graph embeddings’, and ‘language processing’.
* Dim 2 (Vertical): This dimension separates more general business-oriented terms at the top from more specific engineering/technical terms at the bottom. At the top are ‘human’, ‘service industry’, and ‘smart products’, suggesting a focus on end-users and service delivery. Moving down, we find ‘maintenance services’, ‘data mining’, and, at the extreme negative end, ‘language processing’ and ‘language’.
Clusters and Keyword Relationships:
- Cluster 1 (Top Right): A dense cluster towards the top right quadrant includes keywords like ‘human’, ‘service industry’, ‘simulation’, ‘digital storage’, ‘smart products’, ‘servitization’, ‘industrial management’, and ‘product-service systems’. This cluster seems to represent a focus on the *application* of technology within service-oriented industries, potentially including aspects of human-computer interaction, intelligent systems, and new business models (“servitization”). The presence of “simulation” suggests modeling and analysis within this domain.
- Cluster 2 (Top Left): The top left quadrant contains terms like ‘product service’, ‘manufacturing industries’, and ‘decision support systems’. This suggests a cluster related to the production and management aspects of manufacturing and service provision, maybe from the perspective of business strategies.
- Cluster 3 (Bottom Left): Keywords in the bottom left quadrant, such as ‘service-delivery process’, ‘customer satisfaction’, and ‘psmanufacturing companies’, suggest a focus on the core processes of service delivery, customer-centricity, and potentially performance measurement within manufacturing-related companies.
- Cluster 4 (Bottom Right): The bottom right quadrant is populated with keywords like ‘systems engineering’, ‘semantics’, ‘graph embeddings’, and ‘language/language processing’. This cluster represents a more technical and computational facet, potentially related to data analysis, knowledge representation, and AI applications within these fields. The emergence of ‘graph embeddings’ and ‘language processing’ together is interesting, maybe showing the increased use of semantic analysis within these systems.
- Central Keywords: Terms near the origin (0,0), such as ‘industry 4.0’, ‘sales’, ‘ontology’, and ‘maintenance’, can be interpreted as being broadly related to the overall themes but not strongly associated with any specific cluster. These terms could represent bridging concepts or areas of overlap between the different research streams.
Interpretation of Specific Keywords:
- “Industry 4.0”: Its central position suggests it’s a broad umbrella term connecting different research areas within the dataset.
- “Servitization” and “Product-Service Systems”: Their proximity in the top-right cluster reinforces the concept of a shift towards service-oriented business models in manufacturing.
- “Graph Embeddings” and “Language Processing”: Their location in the bottom right points to the rising importance of advanced data analysis and AI techniques in the analysed research area.
- “Customer Satisfaction” and “Service-Delivery Process”: Their presence highlights the importance of operations and marketing in manufacturing/service industries.
Critical Discussion and Further Investigation:
- Variance Explained: The relatively low percentage of variance explained by the first two dimensions suggests that there might be other significant factors or dimensions not captured in this 2D representation. Exploring additional dimensions in the MCA might reveal further insights.
- Parameter Choices: Consider the impact of the parameters used. *`minDegree: 3`* implies that only keywords co-occurring at least 3 times were included. Increasing this value would focus on the most prominent relationships but could also exclude potentially relevant niche topics.
- Database Bias: This analysis is based on Scopus data. Results might differ if the analysis were performed on Web of Science or other databases.
- Temporal Trends: This is a static snapshot. Analyzing the evolution of these keyword relationships over time could reveal emerging trends and shifts in research focus.
- Domain Specificity: The terms “human” and “language processing” are quite generic. Further analysis might reveal more specific sub-topics within these broad areas.
- Stop Words: Depending on the stop words that were removed during the analysis, you might want to consider whether they significantly impacted the results.
In Summary:
This factorial map provides a valuable overview of the key themes and relationships within the Scopus dataset related to keywords in the ‘KW_Merged’ field. The analysis suggests several distinct research streams centered around:
1. Application of intelligent systems and data-driven technologies in service industries.
2. Business strategies around manufacturing/services.
3. Core operations and customer-centricity in the delivery of services.
4. Application of advanced data analysis techniques such as graph embeddings and language processing.
Further investigation, considering the limitations discussed above, will help refine these insights and potentially uncover more nuanced interpretations. Remember to always link these findings back to the specific research questions you are trying to address.
Good luck with your research! Let me know if you have further questions.
Co-citation Network
Overall Structure:
The network appears to have a relatively fragmented structure, with several distinct clusters and a few isolates. This suggests that the research area represented is comprised of several sub-disciplines or distinct lines of inquiry that cite different bodies of literature, and there is not a solid consensus on a theoretical ground between the different communities.
Community Detection (Walktrap Algorithm):
The “walktrap” algorithm has identified several communities, indicated by different colors. This is a good sign; it shows that the algorithm is picking up on distinct groupings of co-cited articles. Let’s break down what we can infer from some of these communities:
- Red Cluster (Top-Left): This appears to be a central cluster, containing important references from Eisenhardt, Goedkoop, and Coreyen. Because the edges are thick, these references are highly interconnected and very important in the literature. Likely core publications in the area.
- Orange Cluster (Left): Closely connected to the red cluster, this smaller group includes Opresnik and Vendrell-Herrero. Given its proximity to the “core” cluster, it probably builds upon or extends the research represented there, perhaps focusing on a specific application or aspect.
- Purple Cluster (Central): This contains Valencia and is connected to multiple other clusters. It may act as a “bridge” between different communities, suggesting it synthesizes ideas from multiple areas or applies a methodology relevant to several sub-disciplines.
- Green Cluster (Bottom-Central): This community appears to contain works of Liu and Chiu. Given that it is connected to Valencia, it could be related to new trends of research.
- Blue Cluster (Right): This is another distinct cluster. Given the high degree of separation with the rest of the clusters, this means that it represent a more specialized sub-topic.
- Pink Cluster (Bottom-Left): This small cluster (Annarelli and Tukker) is isolated. It may signal a relatively new or emerging research area that hasn’t fully integrated with the other established communities, or it could indicate a specific, niche application.
- Brown Cluster (Right): Similar to the pink one, this is another isolated community.
Most Connected Terms (Labelled):
The parameters specified the top 50 nodes were labelled. Those labels give a preliminary insight on important publications in the context of this collection. The size of the labels indicates the number of citations and connections.
- Eisenhardt et al. 1989: This is very prominent in the network, which suggests a seminal reference. Depending on the specific context of your dataset, it could relate to grounded theory methodology, case study research, or innovation management.
- Goedkopp et al. 2016 & Coreyen et al. 2017: It seems that these publications are also relevant.
Recommendations for Further Investigation:
1. Content Analysis: The *most important* step is to *read* the most highly cited papers and those that act as bridges between communities (e.g., Valencia). Understanding the *content* of these papers is crucial to interpreting the meaning of the network.
2. Database Coverage: SCOPUS is a broad database, but it’s worth considering whether the results might differ significantly if the analysis was performed using another database.
3. Keyword Analysis: Correlate this co-citation network with a keyword co-occurrence network to further refine your understanding of the thematic areas represented by each community. Do keywords associated with, say, the “Red” cluster align with the content you find in the Eisenhardt paper?
4. Temporal Analysis: Consider how the network evolves over time. Are certain clusters becoming more dominant? Are new clusters emerging? This will add a dynamic dimension to your interpretation.
Critical Considerations:
- Bias: Co-citation networks reflect citation practices, which can be subject to bias (e.g., self-citation, journal prestige). Be mindful of this when drawing conclusions.
- Level of Analysis: The interpretation of this network should be consistent with the objective of your analysis. Are you interested in understanding the intellectual structure of the field *broadly*, or are you focusing on a *specific* sub-area?
- Data Cleaning: It’s crucial to double-check the data cleaning process. Typographical errors in cited references can create spurious nodes and distort the network.
By combining this network analysis with a deep understanding of the subject matter, you can develop a rich and insightful interpretation of the intellectual landscape of your field.
Historiograph
Overall Observations:
- Temporal Scope: The network spans research from 2015 to 2025 (projected publications), indicating a relatively active and evolving research area.
- Database Source: The data originates from SCOPUS, suggesting a focus on peer-reviewed literature.
- Cluster Structure: The network is divided into distinct clusters (colored nodes), implying different sub-themes or research communities within the broader PSS field.
- Forward Citation Bias: Some 2024/2025 articles appear as terminal nodes, lacking connections, indicating recent publications and the temporal nature of the network.
Cluster-Specific Analysis:
Let’s analyze each cluster chronologically, highlighting pivotal works and thematic trends:
1. Green Cluster (Earliest Focus: 2015 – 2019):
- Key Paper: `lagemann h, 2015: Pss Production Systems: A Simulation Approach For Change Management`. This appears to be a foundational paper within this cluster, using simulation for PSS production.
- Evolution: The cluster builds upon the simulation of change management in PSS production systems.
- Topic: Early modeling and efficiency analysis.
2. Red Cluster (Focus: 2018 – 2025):
- Key Paper: `heinis tb, 2018: Recent Advances And Future Trends In Advanced Prognostics For Smart Machines And Product Service Systems`. This paper likely sets the stage for using prognostics in PSS, perhaps predictive maintenance or performance optimization.
- Pivotal Paper: `wang z, 2021: The Influence Of Dynamic Business Models On Ips2 Network Planning – An Agent-Based Simulation Approach`
- Evolution: Shift towards Dynamic business models, IPS2 (Industrial Product-Service Systems) network planning and agent-based simulation. The cluster extends into topics like aftermarket production systems and quality determinants in digital voice-of-customer.
- Topic: Advanced prognostics, dynamic business models and IPS2 network planning.
3. Blue Cluster (Focus: 2019 – 2024):
- Key Paper: `chen d, 2019: Adaptive Change Management For Industrial Product-Service`.
- Evolution: Focus on adaptive change management in industrial PSS with a focus on requirement elicitation in smart product service systems.
- Topic: Adaptive change management and requirement elicitation in smart PSS.
4. Purple Cluster (Focus: 2022 – 2023):
- Key Paper: `agrawal r, 2022: A Hybrid Crowdsensing Approach With Cloud-Edge Computing Framework For Design Innovation In Smart Product-Service Systems`
- Evolution: The cluster evolves towards user experience in smart tourism products.
- Topic: Crowdsensing, Cloud-edge computing and smart tourism products and services.
5. Orange Cluster (Focus: 2024 – 2025):
- Key Paper: `aeddula o, 2024: A Graph-Based Requirement Elicitation Approach In The Context Of Smart Product-Service Systems`.
- Evolution: Application in smart product service system development.
- Topic: Graph based approaches in smart PSS.
General Interpretation Guidance:
- Citation Strength: Consider the weight or frequency of citations between papers (if available in the underlying data). Stronger links indicate greater influence.
- External Influences: Be mindful that this network represents a specific subset of literature indexed in SCOPUS. The inclusion of other databases (e.g., Web of Science, IEEE Xplore) might reveal different influential papers or trends.
- Network Metrics: Calculate network centrality measures (e.g., betweenness centrality, eigenvector centrality) to identify “broker” papers that connect different clusters or highly influential papers within a cluster.
By examining the titles and publication dates, we can infer the progression of research interests and the emergence of new subfields within PSS. Remember to corroborate these findings with a closer reading of the full-text articles to gain a more nuanced understanding of their contributions.
Collaboration Network
Overall Structure
The network appears to be quite fragmented, consisting of several distinct clusters or communities with relatively few connections *between* these groups. This suggests a research landscape where collaboration is strong *within* specific groups, but less so across different research teams or areas. The absence of a large, central component indicates that there isn’t a single dominant group of researchers connecting everyone in the field (at least, not within this particular dataset).
Communities (Clusters)
The ‘walktrap’ clustering algorithm has identified several communities, visually represented by different colors. Let’s examine some of the key clusters:
- Purple Cluster (Pezzotta G, Pirola F, Sala R): This is a closely knit group. The strong connection between these authors indicates frequent collaboration and possibly a focused research area.
- Red Cluster (Larsson T, Teddula OK, Aeddula): Similar to the purple cluster, this is a well-defined group, with close collaborations. ‘Larsson T’ looks like a central node within this cluster, perhaps indicating a leading role or a coordinating function in their collaborations.
- Orange Cluster (Zheng P, Wang Z, Lix X): A smaller group, but visible. The names suggest this cluster might be located in China.
- Green Cluster (Parida V, Kohtamäki M): Another smaller group.
- Other Smaller Clusters: There are several other very small clusters (e.g., the blue (Alp & Ekuhlenkotter), pink (Zhang & Tang), grey (Belkadi & Bernard) ones). These indicate isolated collaborations or perhaps smaller, newer research initiatives.
Relevance of Most Connected Authors
The nodes are sized proportionally to their degree (number of connections). Therefore, the largest nodes represent the authors with the most collaborations within this dataset. Based on the names, these authors could be key figures in their respective subfields, leading research projects, or working on interdisciplinary research.
Here’s how to interpret the prominence of specific authors:
- Highly Connected Authors (e.g., Pezzotta G, Larsson T, Zheng P, and others in clusters): These authors likely play a significant role in their respective research communities. They may be leading researchers, mentors, or individuals actively involved in multiple collaborative projects. Their work might be central to the topics covered in the analyzed SCOPUS dataset.
- Authors in Smaller Clusters (e.g., Alp Ekuhlenkötter, Belkadi, Zhang): While these authors might be very productive, their collaborations, as reflected in *this specific dataset*, are more limited. It could mean they work in a more niche area, are early-career researchers, or collaborate with individuals not captured in this SCOPUS dataset.
Critical Discussion Points & Further Investigation
1. Inter-cluster Connections: The lack of connections *between* clusters raises questions. Why are these communities so separate? Are they working on completely different aspects of the same broader topic, or are they truly distinct research areas? Further analysis could examine the keywords and abstracts of the publications associated with each cluster to determine the thematic differences and potential overlaps.
2. Database Bias: Remember that this network is based on *SCOPUS* data. Authors who primarily publish in journals not indexed by SCOPUS will be underrepresented. Consider repeating the analysis with Web of Science or other databases to see if the network structure changes significantly.
3. Normalization: The `normalize = “association”` parameter means the edge weights represent the extent to which authors co-author more frequently than expected by chance. This helps correct for differences in publication rates. However, alternative normalization methods could be explored.
4. Temporal Trends: This is a static snapshot. Analyzing collaboration networks over time (e.g., by creating networks for different time periods) could reveal how collaboration patterns have evolved, whether new communities have emerged, and whether there’s increasing or decreasing integration of research efforts.
5. Geographical Considerations: Examining the affiliations of the authors could reveal whether geographical factors influence collaboration patterns. Are the clusters primarily composed of researchers from the same institutions, regions, or countries?
6. Impact of SCOPUS Download Scope: The results are highly dependent on the initial data collection. If the search terms were narrow, it can affect the generalizability of the results.
In summary, this author collaboration network provides a valuable overview of collaboration patterns within the defined SCOPUS dataset. The fragmented structure highlights the presence of distinct research communities, each with its key players. Further investigation is needed to understand the thematic relationships between these communities, the potential influence of database biases, and the temporal evolution of collaboration patterns.
Countries’ Collaboration World Map
Overall Observations:
The map clearly indicates that scientific research output and collaboration are not evenly distributed globally. A significant portion of research is concentrated in specific regions, particularly North America, Europe, and East Asia.
Key Hubs of Scientific Production:
- United States: The US appears to be a major scientific hub, showing a high research output and likely playing a central role in numerous international collaborations.
- China: China shows as another significant hub. The intensity of color suggests a very high research output.
- Europe: Several European countries (e.g., Germany, UK, France, Italy) are visible as significant contributors to global research output.
- Australia: Appears to be another significant contributor
Key International Partnerships (Based on the lines shown):
- China-Malaysia: There is collaborative link between China and Malaysia, as evidenced by a visible red line.
- China-Germany: There is a collaborative link between China and Germany, also evidenced by a red line.
Global Patterns of Collaboration:
- North America & Europe: Likely strong collaboration ties between North America (primarily the USA) and Europe.
- East Asia (China, Japan, South Korea) & Rest of the world: Also show as an important source for collaboration with several countries worldwide.
- Global North vs. Global South: The map shows that North American and European countries may have less collaboration with African countries based on the map presented. This may also apply to other countries in South America.
Considerations & Limitations:
- SCOPUS Bias: Remember that this analysis is based on SCOPUS data. SCOPUS tends to have a broader coverage than Web of Science, but it may still have biases in terms of language, journal coverage, and regional representation.
- Data Normalization: The map’s color intensity is proportional to the *total* number of articles. It doesn’t account for differences in population size or GDP. A country with a large population might have a higher total output simply due to scale, not necessarily a higher per capita research intensity.
- Co-authorship Interpretation: The map uses co-authorship as an indicator of collaboration. While co-authorship is a common metric, it’s important to recognize its limitations. Co-authorship doesn’t necessarily imply deep collaboration or equal contributions from all parties.
- Lines as Binary Indicators: The lines on the map are binary indicators of collaboration (yes/no). They don’t represent the *strength* or *frequency* of collaboration. It would be helpful to know the number of articles behind each connection to better understand the relative importance of each partnership.
- Field Specificity: The analysis doesn’t differentiate between research fields. Collaboration patterns may vary considerably between disciplines (e.g., physics vs. social sciences).
Suggestions for Further Investigation:
- Normalize the data: Calculate research output per capita or per GDP to get a more nuanced view of research intensity.
- Analyze the strength of collaboration: Quantify the strength of collaboration by counting the number of co-authored publications between each pair of countries.
- Examine specific fields: Break down the analysis by research area to identify field-specific collaboration patterns.
- Temporal analysis: Analyze how collaboration patterns have changed over time.
By considering these points and conducting further analysis, you can gain a deeper understanding of international scientific collaboration and its implications for your research area.
