“Business Model*” and “as a service” – Bibliometrix

Overall Collection Scope and Temporal Coverage:

• Timespan: 2002:2025: The data spans a period of 24 years (including the current year, 2025). This is a reasonably long timeframe, which allows for observation of trends and changes in the field. The inclusion of 2025 data suggests either incomplete data for the current year or projections within the database.
• Documents: 1160: You have a collection of 1160 documents. This is a moderate-sized collection. Whether it’s a *large* or *small* collection depends on the specific research area. If you are researching a very niche topic, 1160 documents might be considered a comprehensive set. If your research area is broad, it might represent a more focused selection.
• Annual Growth Rate %: 19.57: This is a *significant* annual growth rate. It indicates that the research area represented by your collection has been experiencing substantial growth in publications over the past two decades. This could be due to increasing interest in the topic, emerging subfields, new methodologies, or increased funding. It warrants further investigation to understand the specific drivers of this growth.

Source Diversity and Productivity:

• Sources (Journals, Books, etc): 754: The documents are sourced from 754 different publications. This suggests a good level of diversity in the publication outlets relevant to your research area. A higher number of sources can indicate a more interdisciplinary field or one where research is disseminated across a wide range of specialized journals. You might want to analyze *which* sources are most prevalent within your collection.
• Document Average Age: 6.87: The average age of the documents is approximately 7 years. This suggests that the collection includes a mix of relatively recent and older publications. It is relevant to explore the literature and build upon existing research, although it is not too aged as to be irrelevant.

Impact and Citation Patterns:

• Average citations per doc: 17.07: On average, each document in your collection has been cited ~17 times. This is a valuable metric, but *its interpretation depends heavily on the research area*. Some fields have inherently higher citation rates than others. You should compare this average to typical citation rates within your specific field to determine if it is considered high, average, or low. The higher the average number of citations per document suggests a greater impact and influence of the collected research.

Content and Keyword Analysis:

* Keywords Plus (ID): 5492; Author’s Keywords (DE): 3112: These numbers indicate the breadth of topics covered within the documents. Keywords Plus are terms automatically generated by the database (SCOPUS), while Author’s Keywords are provided by the authors themselves. The difference in numbers can be insightful:
* Higher Keywords Plus: Might suggest that the database is identifying concepts *beyond* what authors explicitly state, perhaps revealing underlying themes or related areas that authors haven’t directly emphasized.
* Author’s Keywords provide a controlled vocabulary: They give more insights of what the authors themselves believe their works are about.
* Analyzing the *specific* keywords (both types) can reveal the key themes, concepts, and research directions within your collection.

Author Productivity and Collaboration:

• Authors: 3173: A large number of authors have contributed to the 1160 documents. This indicates a wide and diverse community of researchers working in this area.
• Authors of single-authored docs: 139; Single-authored docs: 149: There are 149 single-authored documents. This suggests a presence of independent research efforts alongside collaborative work. Analyzing the content of these single-authored documents might reveal unique perspectives or foundational contributions.
• Co-Authors per Doc: 3.26: On average, each document has ~3 authors. This points to a collaborative research environment. Collaboration is often associated with more complex projects, interdisciplinary research, and greater access to resources.
• International co-authorships %: 21.55: Approximately 22% of the documents involve international collaborations. This suggests a significant level of global engagement in the research area. International collaboration can lead to increased impact, knowledge sharing, and diverse perspectives.

Document Types:

* article: 427; book: 9; book chapter: 110; conference paper: 537; conference review: 34; editorial: 1; retracted: 1; review: 40; short survey: 1: The distribution of document types provides insights into the preferred modes of knowledge dissemination in this field:
* Dominance of Conference Papers (537) and Articles (427): This is a common pattern, suggesting that both peer-reviewed journal publications and conference proceedings are important channels for communicating research findings. A greater number of conference papers compared to articles might indicate a rapidly evolving field where preliminary findings are often presented at conferences before formal publication.
* Presence of Books and Book Chapters: Indicates the existence of more comprehensive and consolidated knowledge syntheses.
* Reviews (40): Suggest that there are efforts to synthesize existing research and provide overviews of specific topics within the field.
* Retracted article: This indicates that a study has been identified to contain error or fraud.
Overall Interpretation & Next Steps:

Based on these statistics, your collection seems to represent a growing, collaborative, and internationally engaged research area with a diverse range of publications. To gain a deeper understanding, consider these next steps:

1. Field-Specific Context: *Crucially*, compare these metrics to typical values within your specific research area. What are the average citation rates for similar publications in this field? Are international collaborations more or less common?
2. Source Analysis: Identify the *most influential* sources (journals, conferences) within your collection. Which journals publish the most frequently cited articles?
3. Keyword Network Analysis: Explore the relationships between keywords to identify emerging themes and research clusters. Use techniques like co-word analysis.
4. Citation Analysis: Analyze the citation network within your collection. Which documents are most frequently cited? Who are the most influential authors? Are there citation bursts indicating breakthrough publications?
5. Trend Analysis: Examine how these metrics have changed over time (from 2002 to 2025). Has the annual growth rate been consistent? Have collaboration patterns shifted?
6. Document type analysis: Explore trends in different types of documents over time.

By combining these statistical insights with a deeper qualitative analysis of the content, you can develop a comprehensive understanding of the research landscape represented by your collection. Good luck!

Annual Scientific Production

Three-Field Plot

Overall Structure and Key Observations

• Central Node (Authors): The ‘AU’ field in the center acts as the focal point. The plot shows which authors are most strongly associated with specific keywords and which references they are citing. The height of the bars in the ‘AU’ column indicates the relative frequency or importance of each author within the analyzed dataset.
• Left Node (Cited References): The ‘CR’ field on the left shows the cited references. The connections (flows) emanating from these references indicate which authors in your dataset are citing them. Thicker flows means more authors are citing the reference.
• Right Node (Keywords): The ‘KW\_Merged’ field on the right shows the keywords. The connections flowing into these keywords indicate which authors are using those keywords in their publications. Thicker flows suggest a stronger association.
• Network Interpretation: By examining the flows between the three fields, we can start to understand intellectual connections and research trends within your dataset. For instance:

Specific Observations and Potential Interpretations

1. Key Authors and their Connections:

* ‘giessmann a’ and ‘reich c’: both are strongly associated with “business models” keyword. In addition, “giessmann a” also shows connection to the “cloud computing” keyword.
* ‘hensher da’: this author has flows to both “cloud computing” and “business model”. The cited references also suggest that this author may also focus on mobility as a service (maas).
* ‘li’ y. Voege ‘t.’: This author is associated with the ‘mobility as a service (maas)’ keyword.

2. Keyword Clusters:

* There seems to be a cluster around “business models”, “cloud computing” and “business model”.
* Another cluster is around “software as a service (saas)” and “web services.”
* Finally, there seems to be a cluster around “mobility as a service (maas)” and “urban transport service”.

3. Citation Analysis:

* ‘vandermerwe s. rada j.’ has a strong citation link to the authors with “business models”.
* The “Mobility as a service (Maas)” cluster has strong reference to “hensher d.a. mulley c. mobility as a service (maas)”.

How to use this for Research Interpretation

1. Identifying Core Literature: Use this plot to identify the most influential papers (highly cited references) in your area. Are there any seminal works that are consistently referenced by multiple authors in your dataset?
2. Mapping Intellectual Lineage: Trace the flows from specific cited references to the authors in your collection. This can reveal who is building upon whose work, and how ideas are evolving.
3. Author Positioning: Examine the connections of specific authors to different keywords. This will help understand their research focus and their contribution to the field. Are they bridging different sub-topics (e.g., connecting ‘cloud computing’ with ‘business model’)?
4. Identifying Research Trends: Look for keywords that are strongly associated with recent publications (check publication years in your dataset). This can highlight emerging areas of research. Are there any shifts in keyword usage over time?
5. Data-Driven Storytelling: Use this plot as a visual aid when presenting your literature review or research findings. It can provide a compelling way to illustrate the relationships between key authors, concepts, and prior research.

Critical Considerations

• Database Bias: SCOPUS has its own coverage characteristics. Results might differ if you used Web of Science, for example.
• Keyword Normalization: The ‘KW\_Merged’ field suggests that some keyword merging/cleaning has been performed. Understanding how this was done is crucial to avoid misinterpretations.
• Granularity: Consider the level of detail appropriate for your analysis. You may need to filter or aggregate the data to focus on the most meaningful relationships.

In summary, this three-field plot offers a valuable overview of the intellectual landscape of your research area. By carefully examining the relationships between authors, keywords, and cited references, you can gain deeper insights into the key themes, influential publications, and emerging trends within your field. Remember to consider the limitations of the data and to interpret the results in the context of your research question.

Most Relevant Sources

Core Sources by Bradford’s Law

Sources’ Local Impact

Sources’ Production over Time

Most Relevant Authors

Authors’ Production over Time

Overall Interpretation:

The plot visualizes the publication history and citation impact of several key authors identified within the Scopus database. Each author’s timeline displays their active years, number of publications per year, and the total citations received by those publications within that specific year. Larger bubbles indicate more publications in a given year, while darker colors signify a higher total citation count per year. This allows us to see not only who is publishing most frequently but also whose work is gaining the most traction.

Individual Author Analysis:

Let’s go through the authors, keeping in mind that the provided data only covers their *top three* most cited-per-year articles, which is important to remember when judging their overall impact:

• GIESSMANN A: Giessmann has an early start with research appearing as early as 2011. The number of publications and citations per year appears moderate, showing consistent but not extremely high impact based on the provided top article list. Top cited article is: DESIGNING BUSINESS MODELS FOR CLOUD PLATFORMS, INFORMATION SYSTEMS JOURNAL, 2016, TCpY 6.3.
• CAI H: Cai H. shows a publishing history from 2009, with a notable publication in 2018 (MODEL-DRIVEN DEVELOPMENT PATTERNS FOR MOBILE SERVICES IN CLOUD OF THINGS, IEEE TRANSACTIONS ON CLOUD COMPUTING, 2018, TCpY 3.9). This is his top cited-per-year article among the list provided. The plot indicates a more distributed publishing pattern across the years.
• BETTONI A: Bettoni A’s publications mainly appear after 2018. Publications appear to be have moderate impact, based on the citations per year metric. A MAAS PLATFORM ARCHITECTURE SUPPORTING DATA SOVEREIGNTY IN SUSTAINABILITY ASSESSMENT OF MANUFACTURING SYSTEMS, PROCEDIA MANUFACTURING, 2019, TCpY 2.4, is the most cited per year work.
• BAUERNHANSL T: Bauernhansl T has a few publications starting from 2020, showing recent activity. However, the citations for their top cited article per year are comparatively low, indicating that while actively publishing, their immediate citation impact (within the year of publication) is still developing. The most cited per year article being DEVELOPMENT OF A METHODOLOGY FOR THE DERIVATION OF TECHNICAL REQUIREMENTS FOR (CYBER-) PHYSICAL PRODUCT-SERVICE SYSTEMS IN SERVICE-ORIENTED BUSINESS MODELS, PROCEDIA CIRP, 2024, TCpY 2, showing recent interest in the field.
• ZHANG Y: Zhang Y’s publications are relatively recent, showing a peak around 2021. The most cited paper is MLCAPSULE: GUARDED OFFLINE DEPLOYMENT OF MACHINE LEARNING AS A SERVICE, IEEE COMPUTER SOCIETY CONFERENCE ON COMPUTER VISION AND PATTERN RECOGNITION WORKSHOPS, 2021, TCpY 9.2.
• WANG Y: Wang Y. shows a couple of publications in the late 2010s and recently 2024. The most cited work being IBM DEEP LEARNING SERVICE, IBM JOURNAL OF RESEARCH AND DEVELOPMENT, 2017, TCpY 3.3.
• REICH C: Reich C. had some significant publications and citations around 2011 (PRIVATE CLOUD FOR COLLABORATION AND E-LEARNING SERVICES: FROM IAAS TO SAAS, COMPUTING (VIENNA/NEW YORK), 2011, TCpY 9.5), suggesting a prominent early contribution, with later work perhaps less impactful (at least within the immediate publication year).
• KETT H: Kett H has a relatively long publication history, showing research published as early as 2010, with publications in 2022. However, the citation count per year is low, which may suggest either a niche focus or a lag in citation accumulation for their work.
• HIDALGO-CRESPO J: Hidalgo-Crespo J. is a recent author in the collection, showing highly cited publications in the last two years: 2023 and 2024. The most cited per year is AN EXPLORATORY STUDY FOR PRODUCT-AS-A-SERVICE (PAAS) OFFERS DEVELOPMENT FOR ELECTRICAL AND ELECTRONIC EQUIPMENT, PROCEDIA CIRP, 2024, TCpY 4.
• RIEL A: Similar to Hidalgo-Crespo J., Riel A. is a newer author, active in the same timeframe and the two co-authored the top three cited articles per year. The most cited per year is AN EXPLORATORY STUDY FOR PRODUCT-AS-A-SERVICE (PAAS) OFFERS DEVELOPMENT FOR ELECTRICAL AND ELECTRONIC EQUIPMENT, PROCEDIA CIRP, 2024, TCpY 4.

Key Observations and Potential Discussion Points:

• Emerging Trends: It looks like “Product-as-a-Service (PaaS)” and related concepts are gaining traction recently, given the activity of Hidalgo-Crespo J. and Riel A.
• Citation Lag: Keep in mind that citation counts are often higher *after* the publication year. Authors with recent publications might see their impact increase in the coming years.
• Publication Venues: Knowing that several publications are in “Procedia CIRP” suggests a conference-driven field, or at least a strong representation in conference proceedings indexed by Scopus. This influences the type of research being captured.
• Data Limitations: As mentioned before, the provided analysis is limited to the top three cited-per-year articles. The general impact and productivity for these authors might be higher if the full range of publications was considered.

Further Research Directions:

• Topic Modeling: Perform topic modeling on the publications to identify specific research clusters and how these authors contribute to different areas.
• Co-citation Analysis: Analyze which authors are frequently cited together to identify intellectual communities and influential works.
• Network Analysis: Create a collaboration network to see who is working together and identify key research groups.
• Longitudinal Study: Extend the timeline of the analysis to see how publication and citation patterns have changed over a longer period.

By considering these points, the researcher can develop a more nuanced and insightful interpretation of the bibliometric data. Remember to always acknowledge the limitations of the data and the specific choices made during the analysis. Good luck!

Authors’ Local Impact

Most Relevant Affiliations

Affiliations’ Production over Time

Corresponding Author’s Countries

Overall Observations:

The plot illustrates the distribution of scientific publications across different countries based on the corresponding author’s affiliation. It differentiates between publications authored solely within a single country (SCP) and those resulting from international collaborations (MCP). This distinction allows us to evaluate both research productivity and the level of international engagement in the field. The data comes from publications indexed in SCOPUS.

Key Findings and Interpretation:

• Leading Countries in Publication Output: Germany (111 articles) and China (76 articles) are the most productive countries in this dataset, significantly outpacing other nations in terms of the number of articles where the corresponding author is affiliated with an institution in that country. This suggests a strong research focus and potentially substantial funding or established research groups in these locations within the specific field represented by the dataset.
• Balance between Domestic and International Research:

* Most countries show a greater proportion of Single Country Publications (SCP) compared to Multiple Country Publications (MCP). This indicates that a significant part of the research is conducted and finalized within national borders. However, the MCP Ratio provides valuable insight into the level of internationalization.
* France has a high MCP % (40.7%), followed by Poland (41.7%), Canada (50%) and United Kingdom (32.5%) suggests a strong inclination towards international collaboration. This might reflect strategic policies promoting international research projects, access to specific expertise or resources not available domestically, or a strong emphasis on global research networks.
* India, although present among the top 10 countries, shows a relatively low MCP percentage (12%), suggesting a stronger focus on domestic research initiatives. Other countries that share this dynamic are China (17.1%), USA (14.6%), Sweden (15.6%), Spain (17.9%), Finland (17.4%) and Portugal (16.7%).

• MCP Ratio as an Indicator of Collaboration Strategy:

* The MCP Ratio offers a valuable metric for comparing the internationalization strategies of different countries. Countries with higher MCP ratios might be actively seeking collaborations to enhance research quality, access diverse perspectives, or address global challenges that require international cooperation. A low MCP Ratio could indicate a preference for domestic funding, infrastructure, or research expertise, potentially driven by national priorities or a focus on local issues.
* Canada stands out with the highest MCP ratio of 50%. This suggests a strong tendency to collaborate internationally. This could be due to various factors such as funding models that encourage international collaboration, a strategic focus on areas where international collaboration is essential, or the nature of research questions being addressed by Canadian researchers.
* Poland stands out with an MCP ratio of 41.7%, showing a strong tendency towards international collaboration. This can indicate a strong emphasis on collaborative research projects and networks.

Potential Discussion Points & Further Exploration:

• Funding Policies: Investigate whether national research funding policies encourage or prioritize international collaborations in different countries. The role of funding and its impact on the extent of collaborative research should be explored.
• Research Focus: The specific research areas or disciplines dominant in each country might influence the need for international collaboration. Some fields naturally require broader global partnerships.
• Geopolitical Factors: Bilateral or multilateral agreements between countries can foster research collaborations.
• SCOPUS Bias: Consider that SCOPUS, while a large database, has potential biases in terms of journal coverage (e.g., favoring English-language publications or journals from specific regions). This bias can influence the apparent research output and collaboration patterns observed.
• Corresponding Author Bias: The analysis is based on the corresponding author’s country. This might not fully capture the extent of collaboration. For example, a paper might involve researchers from multiple countries, but if the corresponding author is from Germany, the entire publication is attributed to Germany.
• Network Analysis: The data could be further analyzed using network analysis techniques to visualize the relationships between countries and identify key collaboration hubs or partnerships.

In summary, this plot provides a valuable overview of research productivity and international collaboration patterns across countries. Analyzing the balance between SCP and MCP, alongside the MCP ratio, allows for a deeper understanding of national research strategies and the drivers behind international scientific partnerships within the field covered by this dataset. Remember to consider the limitations of the data and explore the potential influencing factors mentioned above for a more comprehensive interpretation.

Countries’ Scientific Production

Countries’ Production over Time

Most Cited Countries

Most Global Cited Documents

Most Local Cited Documents

Key Observations & Interpretations

1. High Local Citation, Moderate Global Citation:

* JITTRAPIROM P, 2017, URBAN PLANNING: (LC 37, GC 577, NLC 33.98, NGC 20.43) – This article has the highest local citation count by a fair margin. The global citation count is also substantial. This suggests that this article is highly influential within the specific research domain of your dataset and also has a significant impact on a broader academic audience. The high NLC indicates that, within your dataset, this paper is cited significantly more than average papers of the same age.
* POLYDOROPOULOU A, 2020, TRANSP RES PART A POLICY PRACT: (LC 27, GC 149, NLC 40.5, NGC 6.27) – A similar pattern, though the global citation count is lower than Jittrapiron’s. Still, this paper demonstrates strong relevance to the research topic within your collection. The very high NLC suggests particularly strong local relevance.
* SOCHOR J, 2018, RES TRANSPORT BUS MANAGE: (LC 25, GC 223, NLC 19.69, NGC 6.77) – High local relevance and also a good global citation count.

* *Interpretation:* These articles likely represent core works directly addressing the central themes of your research area. They may contain key methodologies, foundational theories, or significant empirical findings that are heavily referenced by researchers working in this field. The fact that they’re also globally cited implies that their contributions are of broader significance beyond the immediate research area.

2. High Global Citation, Moderate Local Citation:

* XU X, 2012, ROB COMPUT INTEGR MANUF: (LC 7, GC 1694, NLC 12.33, NGC 28.67) – This article stands out with a very high global citation count but a relatively low local citation count.
* *Interpretation:* This suggests the article is highly influential in a broader field (robotics, computing, and manufacturing) but has only moderate relevance within the specific research area defined by your dataset. It might be a seminal work in a related field that occasionally intersects with your research area. The NGC is very high for the article, meaning that in general the paper is very highly cited for its time.

3. Articles with Primarily Local Influence:

* Many of the articles in the list have relatively high local citations but much lower global citations. Examples include: SOCHOR J, 2015, TRANSP RES REC, SMITH G, 2018, RES TRANSP ECON.
* *Interpretation:* These articles may be focused on niche topics, specific regional contexts, or methodologies that are particularly relevant to the research community represented in your dataset, but less broadly applicable or known. They could also be more recent publications that haven’t yet accumulated a large number of global citations.

4. Normalization Matters (NLC & NGC):

* The normalized citation counts (NLC and NGC) provide a more nuanced picture. For example, even if an article has a lower absolute GC, a high NGC indicates that it is highly cited relative to other papers published in the same year. This is particularly important for comparing older and newer publications.
* Consider LUOMA E, 2012, LECT NOTES BUS INF PROCESS: It has low LC and GC, and a low NGC indicating a weak impact in general. However, the NLC is much higher, suggesting that in your collection, it is highly cited, and potentially relevant for your research.

Recommendations for Further Exploration

• Content Analysis: Examine the content of the high-LC/high-GC articles to identify the key concepts, methodologies, or findings that contribute to their influence. How do these articles relate to each other and to the overall themes of your research?
• Citation Network Analysis: Investigate the citation relationships between the listed articles and other documents in your dataset. This can reveal influential clusters of research and key pathways of knowledge diffusion.
• Journal Analysis: Notice the prominent journals in the list (e.g., *Transportation Research Part A: Policy and Practice*, *Research in Transportation Economics*). This can help you understand the key publication outlets in your field.
• Temporal Trends: Analyze the publication years of the cited articles. Are there any shifts in the research focus or influential publications over time?
• Author Analysis: Identify the most frequently cited authors. Are there any leading researchers or research groups that are highly influential in your field?

Critical Discussion Points

• Scope of the Dataset: Remember that the local citation counts are specific to your dataset. The boundaries of your dataset will inevitably influence the apparent “local” influence of articles. Is your dataset representative of the broader field, or is it focused on a specific sub-area or methodological approach?
• Database Coverage: The global citation counts are based on SCOPUS. Consider whether other databases (Web of Science, Google Scholar) might provide different citation counts, especially for articles published in journals not well-covered by SCOPUS.
• Citation Bias: Be aware of potential citation biases. For example, articles published in English-language journals may be more likely to be globally cited. Articles from well-known authors or institutions may also receive more citations.
• Negative Citations: Citation analysis typically doesn’t differentiate between positive and negative citations (i.e., citations that critique or refute a work). A high citation count doesn’t necessarily mean that an article is universally accepted or praised.

By considering these points and conducting further investigation, you can gain a deeper understanding of the intellectual structure of your research field and the key contributions of the most influential publications. Remember to critically evaluate the data and consider the limitations of bibliometric indicators.

Most Local Cited References

Reference Spectroscopy

Overall Plot Interpretation

• Black Line (Cited References by Publication Year): This represents the overall citation activity. A steep upward trend towards the right side of the plot indicates that recent publications are being cited more frequently in the analyzed dataset. The peak of the black line identifies the year(s) containing the most cited publications.
• Red Line (Deviation from 5-Year Median): This line is crucial. It highlights *peak years* where citation frequency significantly exceeds the average citation rate of the preceding 5 years. These peaks represent moments of seminal work, paradigm shifts, or concentrated bursts of influential research within the field. This metric uses a non-centered window, meaning it only looks at the 5 preceding years and does not center the window. This window creates a lag between years with high numbers of citations and the red line, with the most cited papers being those found at the top right of the plot.
• Database: The data was collected from SCOPUS. This is a broad database, but the subject area and search terms used to create the collection will substantially impact the results.

Interpreting Specific Peak Years and Their Key Publications

Based on the provided list of most cited references for the top 10 peak years, we can deduce some key historical and thematic influences in the analyzed field:

* 1967: A clear focus on qualitative research methodologies, particularly Grounded Theory (Glaser & Strauss). This single work appears multiple times, suggesting its foundational and continuing importance. The presence of Zwicky’s “Morphological Approach” indicates an interest in systematic methods for discovery and invention. References to computer utilities also suggest early thinking about computing and information systems.
* Implication: The field seems to have strong roots in qualitative research methods, with Grounded Theory being a particularly influential approach.

* 1981: The most cited document for 1981 is Transportation Research Part B: Methodological. The citation of Fornell & Larcker’s work on structural equation modeling reveals the increasing use of statistical methods for analyzing complex relationships in the field.
* Implication: Quantitative research methods become important

* 1984: Giddens’ “The Constitution of Society” suggests an engagement with sociological theory and its influence on the field. The inclusion of Miles & Huberman’s “Qualitative Data Analysis” reinforces the ongoing importance of qualitative methods.
* Implication: Social theories begin to impact the field

• 1988: A strong emphasis on “Servitization of Business” (Vandermerwe & Rada). This points to a growing interest in service-oriented business models.

* 1991: Emergence of strategic management concepts. Barney’s work on “Firm Resources and Sustained Competitive Advantage” and Grant’s work on “Resource-Based Theory” indicate the adoption of resource-based views in strategic thinking. Ajzen’s “Theory of Planned Behavior” signals an interest in understanding human behavior and decision-making.
* Implication: Strategic theories become important

• 1996: Brynjolfsson and Hitt’s work on the “IT Productivity Paradox” signals concerns and interest in the impact of IT investments on firm performance. Inclusion of Economides “Economics of Networks” suggests that network effects and economics were impactful at this time. Fornell’s American Customer Satisfaction Index” indicated an interest in applying statistical methods to satisfaction indexes.
• 2001: The cited publications, such as Amit and Zott’s work on “Value Creation in E-Business”, focus on the emergence of e-business and the development of the internet.
• 2010: A clear focus on business model innovation, with citations of works by Osterwalder & Pigneur and Chesbrough. This indicates a shift towards studying how businesses create, deliver, and capture value.
• 2017: The papers in 2017 seem to be focused on the circular economy.
• 2020: All of the papers are focused on the new idea of Mobility as a Service (MaaS).

Critical Discussion Points & Further Investigations

• Scope of the Field: The RPYS suggests a field with evolving methodologies and theoretical influences. The shift from qualitative methods to quantitative approaches and the later emergence of business model innovation highlights the dynamic nature of the field.
• Database Bias: Because the data was collected from SCOPUS, it would be good to compare this with data from other databases, such as Web of Science.
• Search Terms: The search terms used in SCOPUS greatly affect the results.
• Modern Focus: The black line shows a focus on current research.

In summary, this RPYS plot provides a valuable historical overview, revealing key influences, methodological shifts, and thematic trends within the field. By examining the cited references in peak years, one can gain a deeper understanding of the intellectual foundations and evolution of this area of research.

Most Frequent Words

WordCloud

TreeMap

Words’ Frequency over Time

Trend Topics

Overall Interpretation:

This plot visualizes the evolution of research topics related to your keyword merged textual field extracted from the SCOPUS database. The plot identifies the top 3 most frequent keywords per year and displays their trends over time. The size of the bubbles indicates the relative frequency of the term within that specific year. The horizontal lines show the interquartile range, representing the spread or variation in the frequency of the term across different documents within the collection for each given year. Terms appearing later in the timeline indicate newer, emerging research areas.

Key Observations and Discussion Points:

1. Early Stage Themes (Early to Mid-2000s):
* The earliest terms, appearing around 2003-2010, center on fundamental concepts like “application service provider,” “software applications,” “applications,” “world wide web,” “computer software,” and “software architecture.” This suggests an initial focus on the basics of software development, the internet, and application delivery.
* “Grid Computing,” “Internet,” and “Information Technology” also emerge during this time, reflecting the rise of distributed computing and the growth of the internet infrastructure.

2. Rise of Cloud Computing and “as-a-Service” Models (2010s):
* Around 2011-2015, we see a clear shift towards cloud computing and related concepts. The terms “software-as-a-service,” “software as a service,” “telecommunication services”, “Business Process,” “Information Services,” “Enterprise resource planning” and “Electronic Commerce” begin to gain prominence. This signifies the growing adoption of cloud-based solutions and service-oriented architectures.
* Terms like “information systems”, “service provider”, “web services”, “cloud computing”, and “application programs” become more prominent, reflecting the development and deployment of cloud-based applications.
* A clear trend towards the XaaS (Everything as a Service) model appears, with “software as a service (saas),” “platform as a service (paas),” and “infrastructure as a service (iaas)” all showing increasing frequency. This indicates a broader shift towards delivering IT resources as services.

3. Business and Economic Considerations (Mid to Late 2010s):
* Around 2017, terms related to business and economic aspects emerge, such as “business modeling”, “sales”, “commerce”, “costs”, “new business models”, and “business model”. This could imply that the focus is shifting toward understanding the business implications and economic impacts of cloud computing and service-oriented models.
* The appearance of “Competition” and “storage as a service (staas)” further emphasizes the competitive landscape of cloud services.

4. Emerging Trends (2020s and beyond):
* In the more recent years (2020 onwards), new topics are coming into focus, including “investments,” “sustainable development”, “mobility as a service,” and “blockchain”. This signals a potential interest in the use of blockchain technology in cloud, and other service models such as mobility.
* More recent trends include sustainability and business model innovation.

Critical Discussion Points and Further Investigation:

• Database Bias: Remember that this analysis is based on data from SCOPUS. Different databases (e.g., Web of Science, IEEE Xplore) might yield slightly different results.
• Keyword Selection: The effectiveness of this analysis depends heavily on the keywords used in the `KW_Merged` field. Understanding how these keywords were selected and merged is crucial. Were they author-supplied keywords, controlled vocabulary terms, or a combination?
• Contextual Understanding: While the plot shows trends, it doesn’t provide the *reasons* behind those trends. You need to delve into the actual publications to understand the specific drivers and motivations.
• Interquartile Range: What does a wide vs. a narrow interquartile range signify for a particular term? A wider range might indicate diverse interpretations or applications of the term, while a narrow range could suggest a more consistent and focused usage.
• Geographical distribution: Where is this research coming from? This can be done by overlaying geographic data of where the publications are coming from.
• Collaboration patterns: Who is working with whom on what topics? You can investigate collaboration patterns and researcher networks using the same data.

In summary, this trend topics plot provides a valuable overview of the evolution of research related to cloud computing, service-oriented architectures, and related technologies. By considering the limitations of the data and further investigating the underlying publications, you can gain a deeper understanding of the key developments and future directions in this field.

Clustering by Coupling

Co-occurrence Network

Overall Structure:

• Network Type: This is a co-occurrence network, meaning nodes (words/keywords) are connected by edges if they appear together within the same articles (in this case, within the title and abstract, since the title is specified as ‘All Keywords network’). The thicker the edge, the more frequently those two keywords co-occurred.
• Normalization: “Normalize: association” means the edge weights reflect the *strength* of association between keywords, adjusted for the individual frequencies of each keyword. A high association indicates that the keywords co-occur *more often than expected by chance*.
• Clustering: The “walktrap” algorithm was used to identify communities (clusters) within the network. These communities represent groups of keywords that are more strongly related to each other than to keywords in other groups. The visual representation confirms three distinct communities identified using the walktrap algorithm, represented by red, green and blue nodes, respectively.

Community Analysis (Topic Identification):

Based on the keywords within each cluster, we can infer the underlying themes or topics:

• Green Cluster (Left): This cluster strongly revolves around Cloud Computing and Software-as-a-Service (SaaS). Key terms include “Software as a Service (SaaS)”, “Cloud Computing”, “Web Services”, “Software as a Service”, “Information Technology”, “Electronic Commerce”, “Application Programs” and “Business Modeling”. This cluster seems to represent the technological and infrastructural aspects of service-oriented business models. This group seems to represent the technological and infrastructural aspects of service-oriented business models. The prominent size of “Software as a Service (SaaS)” and “Cloud Computing” suggests these were central topics in the dataset.
• Blue Cluster (Center/Right): This cluster is centered on Business Models and Related Concepts. It contains terms like “Business Models”, “Innovation”, “Business Model Innovation”, “Industry 4.0”, “Internet of Things”, “Supply Chains”, “Blockchain”, and “Competition”. The connection to “Industry 4.0,” “Internet of Things,” and “Blockchain” suggests a focus on emerging technologies and their impact on business models. The term “Business Models” is central to the cluster. The presence of terms like “economics” and “profitability” suggests a business-oriented perspective.
• Red Cluster (Top): This cluster appears to be focused on Sustainability and Circular Economy. Key terms include “Circular Economy”, “Sustainability”, “Sustainable Development”, “Servitization”, “Service Business Models”, “Product Design”, and “Life Cycle”. This community suggests an exploration of business models driven by environmental concerns and resource efficiency.

Key Terms and Their Relevance:

The size of the nodes corresponds to the number of connections (degree centrality) a keyword has within the network. Larger nodes are more central and represent keywords that co-occur with a greater variety of other keywords.

• “Business Models” (Blue): This is clearly the most central and important term in the entire network. It suggests that the research within this SCOPUS dataset is fundamentally concerned with the study and analysis of various business models.
• “Software as a Service (SaaS)” (Green): This is the second most important term. It indicates that a significant portion of the research in the dataset involves the application of SaaS in different contexts.
• Other Prominent Terms: “Cloud Computing”, “Business Modeling”, “Sustainability”, and “Circular Economy” also have high connectivity. They are key aspects associated to research in the “Business Models” area.

Interpretation Considerations & Potential Research Questions:

• Bridging Concepts: The connections between the clusters can suggest interesting research avenues. For example, the links between the “Cloud Computing/SaaS” cluster and the “Business Models” cluster suggest research exploring how cloud technologies are enabling new business models. The connection between “Business Models” and “Sustainability/Circular Economy” suggests research on sustainable business model innovation.
• Emerging Trends: The presence of terms like “Blockchain,” “Industry 4.0,” and “5G Mobile Communication Systems” within the “Business Models” cluster points to research investigating the impact of these technologies on business model evolution.
• Granularity of Analysis: It’s important to remember this is an “All Keywords” network. Consider creating separate networks based on author-supplied keywords *vs.* title/abstract keywords to see if different perspectives emerge. This can help differentiate between the authors’ intended focus and the broader context of the publications.
• Temporal Analysis: This network represents a static snapshot of the co-occurrence patterns. If your dataset includes publication years, consider creating networks for different time periods to identify how the relationships between these concepts have evolved over time.

In summary, this word co-occurrence network reveals a research landscape focused on Business Models, with strong emphasis on the role of Cloud Computing/SaaS, and an increasing interest in Sustainability and Circular Economy principles. The network provides a valuable overview of the key themes and their interconnections within the SCOPUS dataset, pointing to potentially fruitful areas for further research and analysis.

Thematic Map

Overall Structure and Interpretation

The strategic map is a visualization technique used in bibliometrics to represent the intellectual structure of a field. It plots themes (in this case, keyword clusters) based on two key dimensions:

• Centrality (Relevance degree): Indicates the importance of a theme within the overall research network. Themes with high centrality are highly connected to other themes and represent core areas of research.
• Density (Development degree): Reflects the internal development or maturity of a theme. High density indicates a well-developed and specialized area of research, with strong connections among the articles within that theme.

The map is divided into four quadrants, each representing a different strategic role for the themes:

• Motor Themes (Upper Right): High centrality and high density. These are the dominant, well-developed themes that drive the field.
• Niche Themes (Upper Left): Low centrality and high density. These are specialized areas that are well-developed internally but have limited connections to the broader field.
• Basic Themes (Lower Right): High centrality and low density. These are fundamental themes that are important to the field but may not be as actively researched or specialized.
• Emerging or Declining Themes (Lower Left): Low centrality and low density. These are either new and developing areas or areas that are losing prominence.

Cluster Descriptions and Analysis

Here’s an interpretation of each cluster based on its position on the map, the associated keywords, and the most central articles:

1. Business Models (Motor Theme):

* Position: Upper Right Quadrant (High Centrality, High Density)
* Keywords: “Business models,” “business modeling,” “business model”
* Interpretation: This is a core and well-established theme in your dataset. The high centrality suggests it’s a central concept connecting many other areas of research. The high density indicates a mature and specialized body of literature on business models.
* Key Articles:
* GIESSMANN A, 2013, INTL CONF INFO SYST (ICIS 2013): RESHAPING SOC INFO SYST DES. This article’s high pagerank suggests it is a highly cited or well-connected piece within the business models literature.
* FLOERECKE S, 2018, LECT NOTES BUS INF PROCESS. This article likely explores business process aspects related to business models.
* MLADENOW A, 2015, PROC – INT CONF FUTURE INTERNET THINGS CLOUD INT CONF OPEN BIG DATA, FICLOUD/OBD. This suggests a link between business models and emerging technologies like IoT and cloud computing.

2. Cloud Computing (Motor Theme):

* Position: Upper Right Quadrant (High Centrality, High Density)
* Keywords: “Cloud computing,” “software as a service (SaaS),” “web services”
* Interpretation: Similar to “Business Models,” “Cloud Computing” is another dominant theme in your dataset. It is a core area and it’s being actively researched.
* Key Articles:
* LIU CZ, 2015, INT CONF INF SYST: EXPLOR INF FRONT, ICIS. A central piece in the cloud computing domain.
* LUOMA E, 2018, LECT NOTES BUS INF PROCESS. Focuses on cloud computing within business processes.
* JHANG-LI J-H, 2017, ELECTRON COMMER RES. This indicates research focusing on cloud computing in e-commerce.

3. Cyber Security (Niche Theme):

* Position: Upper Left Quadrant (Low Centrality, High Density)
* Keywords: “Cyber security,” “malware,” “cybersecurity”
* Interpretation: This is a specialized area of research that is well-developed internally but has fewer connections to the other themes in your dataset. Cybersecurity is a mature field with its own distinct community, which might explain its relatively lower centrality in a dataset that also includes business and technological themes.
* Key Articles:
* BOTTAZZI G, 2014, ACM INT CONF PROC SER. The most central piece, likely addressing foundational concepts.
* CHAUHAN PS, 2023, COMPUTER. A more recent article, indicating ongoing research in the area.
* MISHRA S, 2021, COMPUT MATER CONTINUA. Potentially explores cybersecurity within specific computational or material contexts.

4. Circular Economy (Basic Theme):

* Position: Lower Right Quadrant (High Centrality, Low Density)
* Keywords: “Circular economy,” “mobility as a service,” “business model innovation”
* Interpretation: This suggests that circular economy is an important concept in the field, but the research in this area may be less developed or specialized compared to the “Business Models” and “Cloud Computing” themes. The presence of “mobility as a service” and “business model innovation” as keywords suggests that the research focuses on how circular economy principles can be applied to new business models, particularly in the mobility sector.
* Key Articles:
* EHRENHÖFER C, 2012, ANNU SRII GLOBAL CONF, SRII. This article might be a foundational piece, given its older publication date and relatively high pagerank within the cluster.
* OTTERBACH N, 2024, RESOUR CONSERV RECYCL. A very recent article, suggesting a renewed interest or shift in focus within the circular economy research.
* PARK C, 2022, WILEY INTERDISCIP REV ENERGY ENVIRON. Focuses on the intersection of circular economy with energy and environmental concerns.

5. Mobility as a Service (MaaS) (Emerging or Declining Theme):

* Position: Lower Left Quadrant (Low Centrality, Low Density)
* Keywords: “Mobility as a service (MaaS),” “public transport,” “shared mobility”
* Interpretation: This theme appears to be in a relatively early stage of development or potentially declining in prominence within the dataset. It’s not strongly connected to other themes and is not a highly specialized area of research within this context.
* Key Articles:
* RÖHRICH F, 2024, INTL J SUSTAINABLE TRANSP. A recent article, perhaps signaling a potential resurgence or new direction for this theme.
* LIU X, 2019, CICTP: TRANSP CHINA – CONNECT WORLD – PROC COTA INT CONF TRANSP PROF. Highlights the relevance of MaaS in the Chinese transportation context.
* MICHELBERGER F, 2018, TRANSP MEANS – PROC INT CONF. Focuses on the means of transport related to MaaS.

Data-Driven Insights and Critical Discussion

• Core vs. Peripheral Themes: The map clearly distinguishes core themes like “Business Models” and “Cloud Computing” from more peripheral ones like “Cyber Security” and “Mobility as a Service.” This suggests that research in your dataset is heavily focused on the intersection of business and technology.
• Interdisciplinary Connections: The co-occurrence of “mobility as a service” in both the “Circular Economy” and “MaaS” clusters suggests a potential link between these two themes. It is possible to find new research directions at the intersection between Circular Economy and MaaS.
• Temporal Trends: The publication dates of the key articles can provide insights into the evolution of these themes. For instance, the relatively recent articles in the “Circular Economy” cluster suggest a growing interest in this area.
• Limitations: The choice of keywords and parameters (especially `minfreq`) significantly influences the resulting map. A higher `minfreq` would filter out less frequent keywords, potentially simplifying the map but also losing nuanced information. Furthermore, the analysis is based on the “KW_Merged” field from Scopus, implying that the results depend on how keywords are assigned and indexed in that database. The pagerank of the articles might not be representative of their importance within the specific field.
• Next Steps: Further investigation could involve examining the relationships between the clusters in more detail, analyzing the content of the key articles, and exploring how these themes are evolving over time. Consider experimenting with different parameters (e.g., different clustering algorithms, varying `minfreq`) to see how the strategic map changes.

By considering these points, you can effectively use the strategic map to understand the structure of your research field, identify key areas of focus, and guide your future research endeavors.

Factorial Analysis

Overall Structure and Interpretation:

The MCA plot visualizes the relationships between keywords based on their co-occurrence in the Scopus dataset. The position of a keyword reflects its association with other keywords. Keywords closer to each other tend to appear together in the same documents more frequently. The axes (Dim 1 and Dim 2) represent the principal components extracted by the MCA. Dimension 1 explains 41.76% of the variance in the data, while Dimension 2 explains 16.96%. This indicates that the horizontal axis (Dim 1) captures the major differentiating factor among the keywords, and the vertical axis (Dim 2) contributes significantly as well.

Key Observations and Cluster Identification:

Visually, it’s possible to identify some loose clusters or groupings of keywords, although the separation isn’t extremely distinct. Here’s a breakdown of potential clusters and their meanings:

1. Cloud and Service-Oriented Technologies (Top Left): This quadrant includes terms like “digital storage,” “storage as a service (staas),” “5g mobile communication systems,” “business modeling” “platform as a service (paas),” “resource allocatioerce,” “costs,” “information meansysteemt” “cloud computing,” and “sales.” This suggests a cluster around modern IT infrastructure, encompassing both infrastructure-as-a-service and platform-as-a-service, potentially related to data management and processing in the context of emerging technologies like 5G.

2. E-Commerce and Application-Based Services (Bottom Left): Terms like “software as a service (saas),” “electronic commerce,” “application programs,” “web sesofcesare-as-a-service,” and “software as a service” are grouped together. This indicates a theme related to software delivery models, application development, and online business activities.

3. Innovation and Sustainability (Bottom Right): The bottom right quadrant contains “servitization”, “innovation,” “product design,” “sustainable development,” “businessſrcular economy” and “sustainability.” This cluster likely represents research focused on evolving business models towards service-based offerings and environmentally conscious practices. The connection between innovation and sustainability suggests a focus on developing new products and services that address environmental challenges.

4. Business Models and Supply Chains (Center-Right): The terms “mobility”, “supply chains,” and “service business models” form a loosely defined cluster. This highlights the importance of mobile technologies and efficient supply chain management in the context of modern service-oriented business approaches.

5. General Business and Technology Concepts (Center): Terms like “business models,” “internet” appear near the origin. Their central location suggests they are relatively common across the dataset and don’t strongly discriminate between different research themes.

Relevance of Most Contributing Terms:

The terms furthest from the origin on either axis are the most influential in defining the dimensions of the MCA.

• Dimension 1 (Horizontal): The keywords on the far left (“application programs”, “software as a service”, etc.) and the keywords on the far right (“sustainability,” “circular economy”) contribute most to this dimension. This suggests Dim 1 might represent a spectrum ranging from traditional software and IT models to more innovative and sustainability-focused approaches.
• Dimension 2 (Vertical): The keywords at the top (“digital storage”, “storage as a service”, etc.) and the keywords at the bottom (e.g. “application programs”, “web service”, etc.) define this dimension. Dimension 2 seems to separate cloud-based infrastructure and emerging tech from more established or traditional web and application development concepts.

Interpretation Considerations and Further Analysis:

• Parameter Choices: The `minDegree = 26` parameter means that only keywords appearing in at least 26 documents were included. Increasing this value would likely result in a more focused map with only the most prevalent terms. Conversely, lowering it would include more niche keywords but potentially at the cost of clarity.
• Stemming: `stemming = FALSE` means that keywords were not stemmed. Stemming would group together variations of the same word (e.g., “sustain,” “sustainable,” “sustainability”). Applying stemming might consolidate some clusters.
• Database Scope: The results are specific to the Scopus database. Analyses on other databases (e.g., Web of Science) could yield different results due to variations in indexing and subject coverage.
• Further Analysis: To better understand the relationships between these clusters, you could:

* Increase `k.max` to allow for more clusters to be identified by the clustering algorithm.
* Examine the documents associated with specific keywords to gain deeper qualitative insights.
* Compare this map to analyses performed on different time periods to identify trends in research focus.
* Examine the loadings of the keywords on each dimension to get a more precise understanding of their contribution.

In summary: This MCA map provides a valuable overview of the relationships between keywords in the Scopus dataset. It highlights clusters related to cloud technologies, e-commerce, sustainability, and business models. The map’s dimensions reflect underlying trends in the research, with Dimension 1 differentiating between traditional and innovative approaches, and Dimension 2 separating cloud-based infrastructure from application development. Further investigation, guided by these initial observations, will reveal more nuanced insights.

Co-citation Network

Overall Structure:

The network visually displays three distinct clusters or communities, each represented by a different color (green, blue, and red). The size of the nodes appears to be scaled by their degree centrality (number of connections), suggesting that larger nodes are cited more frequently alongside other nodes in the network. Edges between nodes represent co-citation relationships, with thicker edges generally indicating stronger or more frequent co-citation.

Community Analysis:

• Green Cluster: This cluster appears to be centered around the works of “Hensher d.a. -1 Wong y.z.” and potentially related to transportation and/or sustainable mobility based on the nodes near them. Shapiro c. 1999, a bit isolated from the main body of this cluster, might indicate a foundational study in the area that the community has since built upon.
• Blue Cluster: This cluster seems to focus on the works of “Mell p. 2011”, “Benlian a. 2011”, “Turner sueño m.a. 2008”, “Buyya r. 2009” and “Chesbrough h. 2002”, This suggests a focus on cloud computing or related technologies and/or open innovation.
• Red Cluster: The red cluster is centered around “Osterwalder A. 2010,” with secondary hubs being “Teece D.J. 2010”, “Chesbrough h. 2007” “Eisenhardt K.M. 1989” “Yin R.K. 2009” etc. This likely represents a community focused on business model innovation. The size of the Osterwalder node suggests it’s a highly influential and frequently co-cited work in this area.

Most Connected Terms (Prominent Nodes):

• Osterwalder A. 2010: This appears to be the most central and highly connected node in the entire network. This suggests that this work is a foundational piece and frequently cited in combination with other papers within the dataset. Based on the name, it’s likely the popular “Business Model Generation” book/framework.
• Hensher d.a. -1 Wong y.z.: The co-citation suggests that this work is foundational in transport and mobility.

Interpretation and Implications:

1. Interdisciplinary Connections: The presence of edges connecting the clusters, particularly the red and green clusters, suggests interdisciplinary connections between the fields of business model innovation, information systems/cloud computing, and (potentially) transport planning/sustainable mobility. For example, it’s possible that researchers are applying business model frameworks (Osterwalder) to analyze or design cloud-based services or to explore mobility-as-a-service concepts.

2. Dominant Theoretical Frameworks: The co-citation network reveals the core theoretical frameworks and influential publications shaping the research field captured by the dataset. The size of the nodes indicates the relative importance and influence of specific works.

3. Research Foci and Trends: By examining the communities and the connections between them, you can identify the main research themes and emerging trends within the dataset. For example, the relationships between the papers within each community would further elaborate on the specific research questions, methodologies, or empirical contexts being explored.

4. Potential Research Gaps: Areas with fewer connections or under-represented nodes might indicate potential research gaps or emerging areas that warrant further investigation. For example, it is possible to explore research areas or methodologies that potentially create connections between the different cluster.

Further Considerations for Critical Discussion:

• Database Bias: This analysis is based on SCOPUS. How might the results differ if you used Web of Science or another database?
• Search Terms: What keywords or search strings were used to retrieve the dataset from SCOPUS? The results are limited by the scope of the search strategy.
• Time Period: What is the time span of the publications included in the dataset? It would be valuable to observe the community evolution to better understand the impact of different publications across different periods.
• Normalization: Consider that older publications might naturally accumulate more citations. Have you considered any normalization techniques to account for citation age?
• Content Analysis: Conduct a more in-depth content analysis of the key publications within each community to gain a deeper understanding of their specific contributions and relationships.

By critically considering these aspects, you can move beyond a descriptive interpretation of the co-citation network and develop more nuanced insights into the intellectual structure and dynamics of the research area. Good luck!

Historiograph

Overall Structure:

The historiograph shows two distinct clusters based on citation patterns, indicated by different colors (red and blue), which suggests two different research streams. The red cluster is more densely connected, indicating a more cohesive and integrated body of literature. The blue cluster is more isolated.

Cluster 1 (Blue): Mobile Wireless Middleware and Early SaaS Impact (2009-2013)

• Temporal Trend: This cluster represents the earlier phase of research in the field, spanning from 2009 to 2013.
• Pivotal Works: `susarla a, 2009` appears to be the starting point. `luoma e, 2012` and `boillat t, 2013` build upon this foundation.
• Knowledge Development: This cluster seems to be related to Mobile Wireless Middleware. Then, it seems to focus on understanding the impact of the Software-as-a-Service (SaaS) concept on software development and service processes. The title of `boillat t, 2013` suggests application in E-commerce and a focus on interoperability, indicating a move towards practical application of SaaS concepts.
• Interpretation: This cluster represents the initial academic exploration of mobile technologies within the cloud service paradigm. This stream of research likely provided early frameworks and considerations for integrating mobile capabilities with emerging cloud services.

Cluster 2 (Red): Service Brokerage, SOA, and Business Model Evolution (2015-2020)

• Temporal Trend: This cluster represents the more recent and active area of research, largely concentrated between 2015 and 2020.
• Pivotal Works: `jittrapirom p, 2017` seems to be a key node within this cluster, as it has multiple connections. This suggests it’s a central work that connects various ideas and research streams.
• Knowledge Development: This cluster represents the later phase of research in the field, with its focus on service brokerage, SOA-based new business models, the transition to SaaS, and the alignment of software engineering practices.
• Interpretation: This cluster demonstrates a broadening of scope. It moves beyond the technical aspects of SaaS to encompass business model innovation (`polydoropoulou a, 2020`, `karlsson icm, 2020`), strategic management (`polydoropoulou a, 2020`), and Enterprise Application Integration (`reck dj, 2020`). The titles highlight the focus on practical implementation and business value creation, suggesting a shift from theoretical exploration to application-oriented research. `smith g, 2018` also suggests research in this area.

Key Observations and Implications:

1. Evolution of Focus: The research shifts from a primarily technical focus (mobile middleware, early SaaS impact) to a more business-oriented perspective (new business models, value proposition).
2. From Technical Foundations to Business Applications: It demonstrates a natural progression from understanding the technical underpinnings of SaaS to exploring its potential for business innovation and transformation.
3. Service Brokerage as a Connector: The prominent position of `jittrapirom p, 2017` suggests that service brokerage is a crucial concept that bridges the gap between technical implementation and business strategy in the “as-a-Service” domain.

Further Research Directions (based on this limited view):

• Investigate the specific contributions of `jittrapirom p, 2017` to understand why it acts as a central node.
• Explore the connection (or lack thereof) between the mobile middleware/early SaaS cluster and the business model/SOA cluster. Are there any later works that bridge these two areas?
• Analyze the research methodologies employed in each cluster. Are there any differences in approach between the earlier and later phases of research?

Important Considerations:

• Scope of the Data: This analysis is based on a specific Scopus search. The results are sensitive to the search terms and inclusion/exclusion criteria used.
• Citation Bias: Citation networks are susceptible to citation bias. Highly cited works may not necessarily be the most *important* or *impactful*, but rather the most visible.
• Network Visualization Limitations: Historiographs provide a simplified view of complex relationships. Other bibliometric techniques (e.g., co-word analysis, thematic analysis) could provide additional insights.

I hope this analysis helps! Let me know if you’d like me to elaborate on any of these points or explore specific aspects in more detail.

Collaboration Network

Overall Structure

• Disconnected Network: The most striking feature is the highly disconnected nature of the network. Instead of one large, interconnected cluster, we see numerous small groups and even some isolated nodes. This suggests a relatively fragmented research landscape, with authors primarily collaborating within small teams and limited cross-group collaboration within the dataset.
• Small Clusters: Most of the network consists of small, tightly knit clusters, often pairs of authors. This signifies strong collaborative relationships within these groups.
• Limited Centrality: The absence of a single, dominant hub suggests that no single author or group is acting as a major connector or broker within the entire research field represented by the data.

Community Detection (Walktrap Algorithm)

• Community Structure: The Walktrap algorithm identified several distinct communities within the network, indicated by the different colors. This further supports the idea that the research landscape is divided into subfields or thematic areas. The communities (colored nodes) suggest researchers working on similar topics or using similar methodologies.
• Community Repulsion: The `community.repulsion` parameter (set to 0.05) likely influenced the spread and distinctness of the communities. A higher repulsion value would have pushed the communities further apart, potentially highlighting the differences between them even more.

Most Connected Authors (Top 50 Labels)

• Prominent Authors: The labels displayed highlight the most connected authors in the dataset. Focusing on the clusters where these authors reside can tell you something about their research area. For example, `hidalgo-crespo j`and `Riel A` appears to be a central name in one of the clusters.
• Collaboration Patterns: The network configuration around these labels indicates specific collaboration patterns. Closely connected names suggest frequent co-authorship or close working relationships.
• Potential Research Areas: You can investigate the publications of these key authors to understand the specific research topics that characterize each community. This will require external searches since the data set used to generate the network has not been provided.

Interpretation Considerations

• Database Scope: This analysis is limited by the scope of the SCOPUS database.
• Normalization: The use of “association” as the normalization method means that the edge weights (strength of collaboration) are based on the tendency of authors to co-occur on publications compared to what would be expected by chance.
• Parameter Choices: The choice of parameters (e.g., `label.n = 50`, `edges.min = 1`) influences the visualization. You might consider experimenting with different parameter values to see how they affect the network structure and interpretation.
• Isolates Removed: The fact that isolates are removed could mask important information about researchers working alone. You might want to consider keeping them in a future analysis to identify outliers or researchers who could be more integrated into the network.

In Summary

The author collaboration network from SCOPUS suggests a research field characterized by relatively isolated collaborative clusters. No single author or group acts as a central hub, indicating a decentralized and potentially fragmented research landscape. Community detection highlights the existence of distinct subfields or thematic areas. Further investigation of the publications of the most connected authors in each community will reveal more about the specific research topics that define each group. This data-driven interpretation provides a foundation for exploring the dynamics and structure of the research field represented in your dataset.

Countries’ Collaboration World Map

Key Observations and Interpretation:

1. Major Hubs of Scientific Production:

* The map clearly shows the United States and Western Europe (especially countries like Germany, the UK, France) as the most prominent hubs of scientific research output, indicated by the darkest shades of blue. This suggests that these regions are leading in research activity within the dataset’s scope.
* China also emerges as a significant research producer, displayed in dark blue, especially in the coastal part. This underscores the increasing importance of China in global scientific research.

2. Key International Partnerships:

* The thickness and density of the connecting lines indicate the strength and frequency of collaboration. A dense network of lines connects the United States and Europe, indicating very strong and frequent collaboration between these two regions.
* There’s also significant collaboration between China and both the United States and Europe. This highlights the importance of these trilateral partnerships.
* Australia appears to have robust collaborations with the US, Europe, and to some degree, China.
* Noticeable collaboration also seems to exist between the US and South America.

3. Global Patterns of Collaboration:

* A general trend shows that developed nations in North America and Europe, and increasingly, East Asia, are at the center of the global research network. Collaboration tends to be stronger among these regions.
* Many countries in Africa, South America, and parts of Asia, are lighter in color, indicating lower research output, according to this Scopus data. Collaboration lines are also less dense, suggesting less frequent engagement in international co-authorship. This highlights potential disparities in research capacity and international integration.
* The map highlights a potential core-periphery structure in global science, where a few leading countries dominate research output and collaboration, while others participate to a lesser extent.

Critical Discussion Points & Further Investigation:

• Database Bias: The analysis is based on SCOPUS data. It’s crucial to acknowledge that SCOPUS has a bias towards English-language publications and certain journals. This might underrepresent the research output and collaboration patterns of countries with a strong focus on publishing in other languages or regional journals. Consider comparing these findings with analyses based on other databases (e.g., Web of Science, Google Scholar) to assess the robustness of the observed patterns.
• Disciplinary Focus: The observed collaboration patterns might be specific to the disciplines covered in the SCOPUS dataset. Different fields have different levels of international collaboration. It would be insightful to repeat the analysis for specific subject areas to understand if the global collaboration patterns differ across disciplines.
• Authorship Conventions: The map considers co-authorship across *all* authors. It might be interesting to also consider the location of *corresponding* authors. This might give a different perspective on where research is *led* versus where it is *contributed to*.
• Historical Context: These patterns reflect a specific time period covered in your SCOPUS data. It would be valuable to investigate how these collaboration networks have evolved over time. Are there emerging partnerships? Are some collaborations weakening?
• Policy Implications: The findings could inform science policy decisions related to promoting international collaboration. For example, governments might consider initiatives to strengthen research partnerships with countries that are currently less integrated into the global research network.
• Network Characteristics: While the map provides a visual overview, consider supplementing it with network analysis metrics (e.g., centrality, betweenness) to quantify the roles of different countries within the global collaboration network.

In summary: This map provides a valuable overview of global research collaboration patterns, highlighting major research hubs and key partnerships. However, it is crucial to interpret these findings within the context of the SCOPUS database and to consider the limitations of bibliometric analysis. Further investigation, considering database bias, disciplinary focus, authorship conventions, and network metrics, will provide a more nuanced understanding of global scientific collaboration.