The Global Inter-City Collaboration Network of Smart City Research: A Spatial Scientometrics Analysis (2010–2022)

Zsófia Viktória Vida; Balázs Borsi

doi:10.12688/f1000research.179395.1

Home Browse The Global Inter-City Collaboration Network of Smart City Research:...

ALL Metrics

-

Views

-

Downloads

Get PDF

Get XML

Export

▬

✚

Research Article

The Global Inter-City Collaboration Network of Smart City Research: A Spatial Scientometrics Analysis (2010–2022)

[version 1; peer review: 1 approved with reservations]

Zsófia Viktória Vida ¹, Balázs Borsi²

PUBLISHED 29 Apr 2026

Author details Author details

¹ Dept. of Science Policy & Scientometrics, Library and Information Centre of the Hungarian Academy of Sciences, Budapest, 1051, Hungary
² Dept. of Strategy, Entrepreneurship and International Business, ESSCA School of Management, Budapest, 1093, Hungary

Zsófia Viktória Vida
Roles: Conceptualization, Data Curation, Formal Analysis, Funding Acquisition, Investigation, Methodology, Project Administration, Validation, Visualization, Writing – Original Draft Preparation, Writing – Review & Editing

Balázs Borsi
Roles: Conceptualization, Data Curation, Formal Analysis, Methodology, Supervision, Validation, Visualization, Writing – Original Draft Preparation, Writing – Review & Editing

OPEN PEER REVIEW

REVIEWER STATUS

This article is included in the Artificial Intelligence and Machine Learning gateway.

Abstract

Background

Global collaborations play an increasing role in knowledge production, yet geographical proximity remains an important facilitator. The social sciences and sciences exhibit diverse output, citation, and collaboration patterns. This spatial scientometric study investigates the interdisciplinary Smart City (SC) field to determine whether field-specific knowledge regimes shape distinct global collaboration patterns.

Methods

Using metadata from 21,930 Web of Science publications, we employed co-word analysis and Louvain community detection to disaggregate the SC field into technology-oriented and social-oriented meta-domains, focusing on inter-city collaborations across two periods (2010–2017 and 2018–2022). Using affiliations, we identified cities and categorized collaborations across five spatial scales: single-affiliation, intra-city, intranational inter-city, intra-macro-regional, and cross-macro-regional. Analysis variables included number of publications, mean Category Normalized Citation Impact (CNCI), and highly cited papers (Top 5% and Top 10%).

Results

While the centres of output were East Asian hubs, the Global North maintains the highest citation impact in SC domain. Collaborations shifted from localized, single-institution research toward global networks. Inter-city collaborations comprised the majority of highly cited papers, with significant cross-regional links. Single-affiliation research remained high in the social-oriented sub-field, though inter-city collaborations gained prominence, and intra-city collaborations stayed vital. In contrast, in the technology-oriented sub-field intra-city research declined, global links strengthened, and high-impact global collaborations were of great importance.

Conclusions

The two major Smart City sub-fields exhibit typical social science and science collaboration patterns. The technology-oriented sub-field follows a globalized “Big Science” model, while social-oriented research relies more on local context and cognitive proximity. Although geographic distance decay still appears, as intra-city collaboration gives the majority of research output share, there is a tendency of rising importance of global collaboration networks, especially in the case of ‘Top 5%’ highly cited papers.

Keywords

smart city; science mapping; spatial scientometrics; inter-city scientific collaboration, distance decay

Corresponding author: Zsófia Viktória Vida

Competing interests: No competing interests were disclosed.

Grant information: This work was supported by the Hungarian Academy of Sciences [KGYNK2023-46]. I confirm that the funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript
The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.

Copyright: © 2026 Vida ZV and Borsi B. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

How to cite: Vida ZV and Borsi B. The Global Inter-City Collaboration Network of Smart City Research: A Spatial Scientometrics Analysis (2010–2022) [version 1; peer review: 1 approved with reservations]. F1000Research 2026, 15:635 (https://doi.org/10.12688/f1000research.179395.1) First published: 29 Apr 2026, 15:635 (https://doi.org/10.12688/f1000research.179395.1) Latest published: 29 Apr 2026, 15:635 (https://doi.org/10.12688/f1000research.179395.1)

1. Introduction

Cities concentrate economic production, innovation, and knowledge-intensive activities, functioning as central hubs in the global knowledge economy (Balland et al., 2020; Hidalgo et al., 2018; Verginer & Riccaboni, 2021). As scientific production becomes increasingly internationalized and interconnected (Csomós et al., 2020), interurban linkages in knowledge creation intensify. Research on inter-city scientific networks remains underdeveloped compared to spatial economics, where interurban linkages have been studied for decades.

In spatial economics, gravity models constitute a foundational framework for explaining inter-city interactions. These models predict that flows between cities scale with their economic mass (e.g., GDP or population size) and decay with geographical distance (Head & Mayer, 2014). Analogous spatial interaction theory mechanisms operate in scientific systems. Empirical evidence demonstrates geographical bias in citation practices and collaboration networks: researchers disproportionately cite geographically proximate work and are more likely to collaborate with partners located nearby, reinforcing distance-decay effects in knowledge production (Hoekman et al., 2010; Pan et al., 2012; Wuestman et al., 2019).

However, most studies examining inter-city scientific collaboration adopt an aggregate perspective, treating science as a homogeneous domain. Inspired by ‘global city network’ studies (Friedmann, 1986; Sassen, 1989, 2005), this literature maps cities’ positions within global knowledge networks but rarely differentiates between distinct scientific fields (Cao et al., 2023; Maisonobe et al., 2016a; Matthiessen et al., 2010; L. Zhang et al., 2025a). Consequently, limited attention has been devoted to how inter-city collaboration structures vary across specific research domains.

This limitation is particularly relevant in the case of Smart City (SC) research. As an inherently interdisciplinary field, SC scholarship integrates technological innovation with governance and social science perspectives (Mora et al., 2019a). Prior research suggests that disciplinary domains differ in collaboration intensity, team structure, and spatial reach (Fernández et al., 2016; Hall et al., 2018). However, it remains unclear whether and how such differences manifest at the inter-city scale within SC research.

This study addresses this gap by examining global inter-city collaboration patterns in the SC research community. First, building on existing SC research sub-field division (see Rasoulzadeh Aghdam et al., 2025, who distinguished 3 technological, and 1 social oriented emerging SC research topics), we disaggregate the field into technology-oriented and social-governance oriented (in short: social-oriented) sub-fields, and analyse whether their inter-city collaboration structures exhibit distinct geographies. Technological research may cluster around infrastructure-intensive hubs and social-oriented research may exhibit stronger local embeddedness. Second, extending the city-level mapping approach of Matthiessen et al. (2010), we identify global patterns of output and impact at the city-level. By incorporating disciplinary diversity – the technology-oriented and social-oriented sub-fields) into the analysis of inter-city scientific networks, this study advances spatial interaction and global city network research, demonstrating how field-specific knowledge regimes shape the geography of academic collaboration.

2. Literature review

2.1. The spatial organization of academic knowledge production

Scientometrics is the quantitative study of science and scholarly communication (van Raan, 2019). Spatial scientometrics is a specialized sub-field, addressing explicitly the spatial and geographical aspects of scientific activities, focusing on how regional concentration or distance influence the creation and impact of scientific knowledge (Frenken et al., 2009a; Frenken & Hoekman, 2014; Gao, 2022).

Spatial interaction models and agglomeration economies have long conceptualized cities as knowledge agglomerations (Fujita et al., 1999). In this view, proximity in urban areas directly impacts economies through more efficient matching and sharing of inputs, boosted by knowledge spillovers (Autant-Bernard et al., 2007; Duranton & Puga, 2004). These knowledge spillovers originate from the spatial concentration of R&D and universities in the cities, highlighted by studies on article production (Csomós, 2018).

These insights have been extended to scientific collaboration and citation networks. Spatial scientometrics research demonstrates that co-authorship ties and citation flows exhibit systematic spatial structures (Frenken & Hoekman, 2014). In line with spatial interaction theory, a gravity-like logic emerges, consistent with formal gravity model specifications (see e.g. Head & Mayer, 2014): collaboration intensity increases with urban research mass (e.g., institutional scale, publication output) and decreases with geographical distance (Csomós et al., 2020; Hoekman et al., 2010; Pan et al., 2012; Wuestman et al., 2019). Distance-decay effects persist in the digital era: urban scale and geographical proximity remain significant predictors of both collaboration probability and scientific impact.

Nevertheless, most large-scale analyses treat science as a homogeneous system. Inter-city collaboration and citation structures are typically examined in aggregate form, despite evidence that disciplinary domains differ substantially in collaboration practices, team size, citation behaviour, and internationalization patterns. Aggregation may therefore obscure important field-specific spatial dynamics.

2.2. Inter-city scientific networks and the global city

Global city network research emphasizes interurban linkages as constitutive elements of global urban hierarchies (Derudder & Taylor, 2020; Friedmann, 1986; Sassen, 1989, 2005; Taylor, 2001). Within this framework, cities derive their global positioning from their integration into transnational flows of capital, services, and increasingly, knowledge.

Applied to science, this perspective highlights how cities occupy differentiated positions within global knowledge networks. Empirical studies map inter-city co-authorship networks to identify central hubs, peripheral nodes, and hierarchical structures in scientific production (Cao et al., 2023; Maisonobe et al., 2016a; Matthiessen et al., 2010; L. Zhang et al., 2025a). These studies consistently reveal pronounced core-periphery configurations, with a limited number of global cities dominating collaboration intensity and citation impact.

Core-periphery structures in scientific collaboration networks reflect gravitational mechanisms similar to those identified in spatial economics: large, research-intensive cities attract disproportionate collaboration flows, while peripheral cities exhibit weaker connectivity. Nonetheless, these analyses rarely differentiate between scientific fields. The implicit assumption of structural uniformity across disciplines may mask substantial variation in how different academic disciplines organize inter-city linkages. If disciplinary areas differ in infrastructure dependency or collaboration norms, then global urban hierarchies in science may not be invariant across fields.

2.3. Thematic heterogeneity and field-specific collaboration patterns

A growing body of research demonstrates that disciplinary heterogeneity shapes collaboration structures. The concept of cognitive proximity (Boschma, 2005a) suggests that collaboration likelihood depends on shared knowledge bases and interpretative frameworks; thus excessive cognitive distance can hinder cooperation (Bu et al., 2025). Disciplinary areas differ in their scholarly impact and the cross-fertilization from other disciplines (Yan, 2016; Y. Zhang et al., 2025b). Team size, collaboration intensity and cross-border reach also differ systematically between fields (Fernández et al., 2016; Hall et al., 2018; Wuchty et al., 2007). Two interrelated mechanisms are particularly relevant for understanding field-specific spatial patterns.

First, infrastructure intensity varies across domains. Engineering and technological disciplines are dependent on critical research infrastructure (see the nanotechnology case in Sandip, 2006), whereas social science, planning, and governance research exhibit more divergent, interpretive knowledge territories that reflect greater context-sensitivity (Becher & Trowler, 2001). The infrastructure requirements favour concentration of technological domains in well-resourced metropolitan research hubs, reinforcing agglomeration dynamics through technology, whereas context sensitivity may anchor social sciences more strongly to specific places and encourage regionally embedded collaboration structures.

Second, collaboration norms and academic cultures vary. Disciplines differ in typical team size, authorship conventions, and incentives for international collaboration (Wuchty et al., 2007), and the spatial reach of their networks is likely influenced.

Despite recognition of disciplinary heterogeneity, systematic cross-city evidence remains limited. Studies occasionally compare broad field categories and document variation in distance sensitivity and collaboration intensity. For example, Abramo et al. (2020) examined the effect of geographic distance on knowledge flows, showing differences between social sciences, humanities, and (natural, engineering, medical, agricultural) sciences. However, few investigations analyse whether gravity-like dynamics differ structurally across different domains at the inter-city level.

Taken together, epistemic regimes and geography are intertwined. Technological domains may exhibit stronger hierarchical concentration in global hubs due to infrastructure dependency and large-scale collaboration norms. Social-oriented research may demonstrate stronger local embeddedness and higher sensitivity to geographical distance. Whether such differences manifest within interdisciplinary domains remains empirically underexplored.

2.4. Smart City (SC) research: An interdisciplinary knowledge domain

SC research provides a particularly suitable case for examining field-specific spatial dynamics within a shared thematic umbrella. Although relatively recent, the field has evolved substantially. Early contributions emerged in architecture and social science journals, while the mid-2000s witnessed rapid expansion driven by engineering and computer science perspectives emphasizing ICT-enabled urban solutions. More recently, governance, planning, and sustainability perspectives have gained prominence, although technological orientations remain influential (Bibri & Krogstie, 2020; Ricciardi & Za, 2015; Shang et al., 2020; Zhao et al., 2021). Overall, SC scholarship integrates technological innovation with governance and social science perspectives (Mora et al., 2019a).

This internal differentiation suggests the co-existence of distinct disciplinary orientations within the same overarching field. Technology-oriented strands may be more infrastructure-intensive and globally connected through large research hubs, while social-oriented strands may be more context-dependent and locally embedded. If so, their inter-city collaboration structures may differ in hierarchical configuration, spatial reach, and sensitivity to geographical distance.

We have therefore identified the following gap. Science is demonstrated to be spatially structured, and inter-city scientific networks have been mapped extensively. However, these analyses largely aggregate across disciplines. SC research offers a natural test case to assess whether disciplinary heterogeneity delineates into distinct inter-city geographies within a single interdisciplinary domain. In particular, it allows examination of whether gravity-like dynamics – urban research mass effects (attraction to places) and distance decay (collaboration decreasing with space) – vary between technology-oriented and social-oriented strands.

Accordingly, this study addresses the following research questions from a spatial scientometric perspective:

RQ1: How is global SC research distributed across cities, and what are the patterns of scientific production, citation impact, and inter-city collaboration?

RQ2: Do technology-oriented and social-oriented thematic domains exhibit distinct scientific production, citation impact, and inter-city collaboration?

RQ3: How does geographical distance affect inter-city collaborations across the two thematic domains in SC research?

By addressing these questions, this study provides a multi-dimensional mapping of the global SC research landscape, uncovering the hidden spatial and thematic logic that governs international scientific collaboration.

3. Data and methods

3.1. Data retrieval and scope

To examine global SC research, metadata were retrieved from the Web of Science (WoS) Core Collection. The search strategy was informed by Subject Category (SC field in WoS) definitions and established conceptual frameworks (Cocchia, 2014; Albino et al., 2015; Eremia et al., 2017), as detailed in Table 1. To maintain consistency, the selection was restricted to English-language journal articles and proceedings papers that were published between 2010 and 2022. After excluding records with missing publication years, a final set of 25,782 publications was retained.

Table 1. Smart City (SC) research search profile.

Search profile

Hits in web of science

(TS = (“intelligent city” OR “knowledge city” OR “ubiquitous city” OR “sustainable city” OR “digital city” OR “smart city” OR “future city” OR “eco city” OR “eco-city” OR “compact city” OR “liveable city” OR “digital city” OR “innovative city” OR “green city”) OR TS = (“intelligent cities” OR “knowledge cities” OR “ubiquitous cities” OR “sustainable cities” OR “digital cities” OR “smart cities” OR “future cities” OR “eco cities” OR “eco-cities” OR “compact cities” OR “liveable cities” OR “digital cities” OR “innovative cities” OR “green cities”) OR TS = ((urban OR city OR cities) AND (“smart econom*” OR “smart people” OR “smart govern*” OR “smart mobility” OR “smart environment*” OR “smart living”))) AND DT = (“Article” OR “Proceedings paper”) and English (Languages)

28444

3.2. Data processing

A primary technical requirement of this study was to isolate the technology-oriented and social-oriented sub-fields within the SC domain to allow for a comparative spatial analysis.

For science mapping, we employed co-word analysis (Callon et al., 1983) as a thematic-detection technique. While methods such as bibliographic coupling (Kessler, 1963; Yan & Ding, 2012) or co-citation (Small, 1973) exist, co-word analysis is the most effective tool for identifying the conceptual content of a field (Zupic & Čater, 2015).

For co-word analysis, the fields DE (Author-Defined Keywords, reflecting the researchers’ primary focus), and ID (Keywords Plus, algorithmically generated by WoS) were used and standardised, using the Keyword Standardisation and Restructuring (KSR) procedure (Borsi et al., 2025). This involved systematic expert-supported manual cleaning in order to merge synonyms and restructure overlapping terms to improve the analytical potential.

Low-frequency keywords (n < 20) were removed, and 772 terms in 22,190 articles remained. The co-occurrence matrix was normalised using Salton’s Cosine coefficient (Salton & McGill, 1983; Hamers et al., 1989) and subsequently clustered using the Louvain community detection algorithm (Blondel et al., 2008), which maximizes modularity and creates non-overlapping clusters.

We divided our temporal range into two periods: a rapid publication explosion period (2010–2017) and a stabilized output period (2018–2022). For the two periods, we got respectively 5 and 6 clusters. In both cases, the largest cluster includes sustainability-related content, with multiple dimensions of sustainability (social, environmental, the social-oriented cluster) while the remaining 4 (2010–2017) and 5 (2018–2022) clusters have a strong technological focus, such as IoT, big data, remote control, and sensors (the technology-oriented cluster) (Figure 1 A, B).

Figure 1. Wordclouds of Keywords of Clusters in Smart City Domain.

A. wordclouds of clusters between 2010-2017. B. wordclouds of clusters in 2018-2022. Note: Larger words indicate higher frequencies within their respective thematic clusters. For the keyword cloud visualizations, minimum keyword occurrence was set at 20.

Thus, following the framework by Bibri and Krogstie (2017), the resulting clusters were aggregated into two meta-domains: a technology-oriented cluster and a social-oriented clusters (Figure 2A--D). This classification served as the basis for all subsequent comparative analyses. Creating wordcloud figures (Figure 1, Figure 2) we used ‘ggplot2’ (Wickham, 2016) and ‘ggwordcloud’ (Le Pennec & Slowikowski, 2024) R Packakes.

Figure 2. Wordclouds of Keywords for social- and technology-oriented Smart City clusters

A. Wordclouds of social-oriented smart city cluster, 2010–2017 (green). B. Wordclouds of technology-oriented smart city cluster, 2010–2017 (orange). C. Wordclouds of social-oriented smart city cluster, 2018–2022 (green). D. Wordclouds of technology-oriented smart city cluster, 2018–2022 (orange). Note: Larger words indicate higher frequencies within their respective thematic clusters. For the keyword cloud visualizations, minimum keyword occurrence was set at 20.

The second phase of data processing involved extracting geographical information from author affiliations. City names were identified and isolated for 5,907 publications in the first period (2010–2017) and 16,023 in the second (2018–2022).

To analyse the evolution of collaborative distance, publications were classified into five spatial collaboration categories reflecting increasing geographical reach:

• Single affiliation (no external collaboration).
• Intra-city collaboration (multiple affiliations within the same city).
• Intranational inter-city collaboration (different cities within the same country).
• Intra-macro-regional collaboration (cross-border within the same macro-region).
• Cross-macro-regional collaboration (inter-continental/cross-macro-regional).

Scientific output was measured by aggregating publication counts to the city level to map globally the distribution of SC knowledge. In this article, we did not use fractional counting for aggregation. To analyse the distance effect, we aggregated research output for the five spatial collaboration categories mentioned above. Scientific impact was measured with several variables. For revealing citation patterns in RQ1 and RQ2, we used Category Normalised Citation Impact (CNCI) from Clarivate’s InCites database. This allows for an unbiased comparison of impact regardless of field-specific citation norms. We calculated the aggregated mean values of CNCI for each city by taking the arithmetic average of the CNCI values of papers produced in the city. We presented scientific output and impact between the two periods 2010–2017 and 2018–2022 on maps. We applied the following Rpackages for creating the maps: ‘ggplot2’ (Wickham, 2016),‘sf’ (Pebesma, 2018; Pebesma & Bivand, 2023), for base map layers (coastlines and country borders) we used from Natural Earth via the ‘rnaturalearthdata’ (South et al., 2024) and ‘rnaturalearth’ package (Massicotte & South, 2026).

Figure 3. Output (number of publications) and impact (mean Category Normalized Citation Index, CNCI) of Smart City (SC) Research.

A. Output, blue, 2010–2017. B. output, blue, 2018–2022. C. Impact, red 2010–2017; D. Impact, red 2018–2022.

Source of base map layers (coastlines and country borders) we used from Natural Earth via the ‘rnaturalearthdata’ (South et al., 2024) and ‘rnaturalearth’ package (Massicotte & South, 2026).

Figure 4. Scientific Output-Impact change between 2010–2017 to 2018–2022 in SC research.

Note: Based on Matthiessen et al., 2010, Figure 6. Output change: percentage change of publication output between period 1 and period 2; Impact change: percentage change of mean Category Normalized Citation Index (CNCI) between period 1 and period 2.

Figure 5. Output (number of publications) and Impact (mean Category Normalized Citation Index, CNCI) of Smart City (SC) social-oriented and technology-oriented sub-fields.

A. Output, 2010–2017, green: Social-oriented SC sub-field; blue: Technology-oriented SC sub-field, 2010–2017. B. output, 2018–2022. green: Social-oriented SC sub-field; blue: Technology-oriented SC sub-field; C. Impact, 2010–2017; green: Social-oriented SC sub-field; orange: Technology-oriented SC sub-field, 2010–2017. D. Impact, 2018–2022. green: Social-oriented SC sub-field; orange: Technology-oriented SC sub-field. Source of base map layers (coastlines and country borders) we used from Natural Earth via the ‘rnaturalearthdata’ (South et al., 2024) and ‘rnaturalearth’ package (Massicotte & South, 2026).

For the investigation of distance decay, we compared scientific output and impact among the five spatial collaboration categories from single affiliation to cross-macro-collaboration. As the outcome was measured with a volume, the percentage of publications in each spatial category, we used a similar volume for comparison, which expresses impact. We chose the percentage of highly cited papers ‘Top 5%’ and ‘Top 10%’ for all spatial categories. This kind of approach was also used in Csomós et al. (2020), using Top 1%. As this article is not focusing on the excellence of science, but rather describes a specific SC research field, we chose ‘Top 10%’ and ‘Top 5%’ indicators. Using both values enables us to find the connection between spatial collaboration categories, output, and strengthening impact. Finally, following the work of Figure 6 of Matthiessen et al. (2010), we mapped the relative change in output and impact to track the evolving positions of cities within the global SC hierarchy. We used the categories found in Matthiessen et al. (2010), and the indicators of output and citation change; since it was not described in detail which kind of citation indicator was used, we chose the mean CNCI for measuring impact on the diagram. For visualization, 49 cities with at least 5 publications in both sub-fields in both time periods were selected. Figure 4, Figure 6, Figure 7, Figure 8 We used ‘ggplot2’ (Wickham, 2016) and ‘ggrepel’ (Slowikowski, 2024) R Packakes for the diagrams in this study.

Figure 6. Scientific output and impact change from 2010–2017 to 2018–2022 in the social-oriented sub-field of SC research.

Note: Based on Matthiessen et al. (2010), Figure 6. Output change: percentage change of publication output between period 1 and period 2; Impact change: percentage change of mean Category Normalized Citation Index (CNCI) between period 1 and period 2.

Figure 7. Scientific Output-Impact change between 2010–2017 to 2018–2022 in social-oriented sub-field of SC research without the outliers.

Note: outliers: Guangzhou, Prague, and Sydney. Figure 7 based on Matthiessen et al. (2010), Figure 6. Output change: percentage change of publication output between period 1 and period 2; Impact change: percentage change of mean Category Normalized Citation Index (CNCI) between period 1 and period 2.

Figure 8. Scientific Output-Impact change between 2010–2017 to 2018–2022 in the technology-oriented domain of SC research.

Note: Based on Matthiessen et al., 2010, Figure 6. Output change: percentage change of publication output between period 1 and period 2; Impact change: percentage change of mean Category Normalized Citation Index (CNCI) between period 1 and period 2.

This rigorous data processing phase transformed a heterogeneous set of 21,930 articles into a structured, dual-track dataset defined by both the representative knowledge regimes (the two sub-fields of SC research) and geographies across the two time periods. This structured dataset provides the necessary foundation to address the research questions.

4. Results

4.1. The global distribution of SC research (RQ1)

Most SC publications concentrate in East Asian and European hubs. India and the Middle East also have relevant output, especially in the second period ( Figure 3A, 3B). Over the two periods, the number of cities involved in SC research is increasing (1,566 and 2,836 cities, respectively in the first and second period). At the same time, a spatial concentration of the research output is observed, with hubs preserving their leading positions. If only cities with at least 10 papers in both periods are considered, their number is more stable (721 and 778 respectively, in the two periods).

The impact of publications exhibits further dynamics. The mean CNCI values for cities demonstrate that while East Asia is strong in research output, significantly higher impact values are located in Europe and in the United States. Overall, the global North holds a strong position in the field of SC research ( Figure 3C, 3D).

As demonstrated in Figure 3, the global state of SC research is illustrated across two periods. Figure 4 presents the percentage change in these dimensions, indicating future trends. There are 49 cities that have at least 5 publications in both SC sub-fields (social-oriented and technology-oriented clusters), in both time periods (n = 49). Figure 4 represents these based on Matthiessen et al. (2010), Figure 6. Cities are categorised by their rate of change of output and impact. Following the nomenclature of Matthiessen et al. (2010), the following groups of cities are distinguished:

• ‘Hot spots’ (red): both scientific output and impact have seen a growth of more than 25% above the mean value of both variables.
• ‘Volume growth but loss in reputation’ (green): only scientific output has grown more than 25% above the mean value.
• ‘Focus on success’ (yellow): only the impact value has grown more than 25% above the mean value.
• ‘Black holes’ (blue): both output and impact values have fallen more than 25% below the mean value.
• ‘Neutral’ (gray): the values of the variables did not exceed a 25% change to the mean.

‘Hot spot’ cities are few; only Sydney, Guangzhou, Nanjing, Shenzhen, and Wuhan belong to this category. The ‘Focus on success’ city group is also remarkable. Despite their relative decline in output, these cities are gaining influence. Apart from the ‘Black holes’, most European cities are in this group. ‘Volume growth but loss in reputation’ cities are geographically diverse, involving cities from Brazil, Iran, China, Singapore, South Korea, and the United Kingdom.

Regarding inter-city collaborations, a decrease is observed in the rate of publications produced in intra-city collaboration over the two time periods: from nearly two-thirds to half; inter-city collaboration is rising ( Table 2). For the highly cited papers, a dramatic increase of inter-city collaborations emerges for the second period. There is a difference between the ‘Top 5 %’ and ‘Top 10%’ highly cited publications: the more highly cited a paper is, the higher the rate of inter-city collaborations.

Table 2. The share of intra- and inter-city collaboration based on research output and highly cited papers in SC field.

	Output share %, 2010–2017	Output share %, 2018–2022
No collaboration or intra-city collaboration	66.1	50.9
Inter-city collaboration	33.9	49.1
Total	100	100

	Top 10% cited pubs share %, 2010–2017	Top 10% cited pubs share %, 2018–2022
No collaboration or intra-city collaboration	58.06	37.47
Inter-city collaboration	41.93	62.53
Total	100	100

	Top 5% cited pubs share %, 2010–2017	Top 5% cited pubs share %, 2018–2022
No collaboration or intra-city collaboration	56.6	34.37
Inter-city collaboration	43.39	65.63
Total	100	100

4.2. The global distribution and comparison of SC sub-fields: technology-oriented vs. social-oriented (RQ2)

The two sub-fields of SC research exhibit divergent trends. The technology-oriented cluster accounted for two-thirds of all papers. We observed an increase in social-oriented papers in Europe and partially in Latin America in the second period; however in the North American region and in Asia the share of this sub-field declined marginally. The research output of global SC hubs covers both SC sub-fields. Within India and Middle East an increase in the number of papers can be attributed to the technology-oriented sub-field (see Figure 5A and 5B). Between 2010–2017, 803 cities participated in the social-oriented sub-field, while 1,303 in the technology-oriented sub-field. In the subsequent period (2018–2022), the number of cities in the social-oriented sub-fields are 1,744, and in the technology-oriented sub-field 2,405. The global north plays a remarkable role in terms of the impact of SC research. The impact of the Social-oriented sub-field is relevant in both periods in Europe and U.S., and in the second period it also appears in Japan and in Australia. China’s research impact is evident in technology-oriented fields for the second period ( Figure 5C, 5D).

Focusing on output and impact change between the two periods, in the case of the social-oriented sub-field, three cities demonstrate extreme positive change: Prague in impact, Guangzhou in output, and Sydney in both. Along with Sydney, we find Paris, Singapore, and Tehran in the ‘Hot spot’ category. Among ‘Volume growth but loss in reputation’ cities belong to a wide array of geographical locations ( Figure 6).

Zooming in the diagram without Prague, Sydney, and Guangzhou (the outliers) in the ‘Focus on success’ group, we mostly find European cities along with Los Angeles and Santiago. Important research centres such as Beijing or London belong to the ‘Black hole’ group ( Figure 7).

In the case of the technology-oriented sub-field of the SC domain, we find no outliers. In the ‘Hot spots’ group, Hong Kong and Nanjing have the greatest positive change of output, and Guangzhou has the highest impact change. In the ‘Volume growth but loss in reputation’ group Manchester, Seoul, Jinan, Singapore, Tehran and Atlanta are found. Most of the cities in the ‘Focus on success’ and ‘Black hole’ groups are located in Europe ( Figure 8).

Comparing Figures 7 and 8, 14 cities are in the same group in both sub-fields. Most of them are in the ‘Black holes’ category, except for Sydney (‘Hot spot’), Seoul (‘Focus on success’), and Santiago (‘Volume growth but loss in reputation’).

4.3. Geographical distance effect on inter-city collaboration networks in SC research (RQ3)

Our findings confirm the trend that scientific collaborations are becoming increasingly common. The share of papers with only one affiliation sharply decreased between the 2010–2017 and 2018–2022 periods, and a higher proportion of articles originates from inter-city collaborations. However, distance still matters, since the majority of publications emerged from intra-city collaborations. Between the two periods, the proportion of articles increased in every inter-city category, and the rate of increase was higher in categories covering greater distances. Therefore, there is a marked transition from localised research to globalised networks. When inter-city collaborations cross national borders, cases between macro-regions are more common than those within them. At the same time, the intranational inter-city networks show a consistent growth for the second period in both output and high impact share. In the case of highly cited papers, the geographical distance decay effect is weakening. While intra-city collaborations between 2010–2017 were 46.36% of the ‘Top 5%’ most cited publications, this share has diminished in the second period to 29.49%. In contrast, cross-macro-regional collaborations represent the largest rate of highly-impact papers, contributing 35.55% of the ‘Top 5%’ cited papers between 2018 and 2022. It is also striking that, in the first period, the distribution of output and highly cited papers across the collaboration scale shows similar proportions, whereas in the second period, the differences between categories are much more pronounced for both output and highly cited papers ( Table 3).

Table 3. The share of output by regional collaboration categories in SC field.

	Output share %, 2010–2017	Output share %, 2018–2022
Single affiliation	13.7	8.2
Intra-city collaboration	52.4	42.7
Intranational inter-city collaboration	13.4	17.8
Intra-macro-regional collaboration	7.2	9.9
Cross-macro-regional collaboration	13.2	21.3

	Top 10% cited pubs share %, 2010–2017	Top 10% cited pubs share %, 2018–2022
Single affiliation	10.52	5.22
Intra-city collaboration	47.54	32.25
Intranational inter-city collaboration	12.94	17.36
Intra-macro-regional collaboration	9.08	12.95
Cross-macro-regional collaboration	19.91	32.22

	Top 5% cited pubs share %, 2010–2017	Top 5% cited pubs share %, 2018–2022
Single affiliation	10.24	4.88
Intra-city collaboration	46.36	29.49
Intranational inter-city collaboration	12.53	16.94
Intra-macro-regional collaboration	9.7	13.14
Cross-macro-regional collaboration	21.16	35.55

Comparing the two sub-fields ( Table 4), the proportion of single affiliation articles is significantly higher, nearly triple (24.52%) in the social-oriented SC sub-field than the 8.73% share of the technology-related SC sub-field in 2010–2017. In the case of output, including highly cited papers in single affiliation category has significantly higher values than in the technology-oriented sub-field in both time periods. Besides, a decreasing share of single-institution work appears in both sub-fields by 2018–2022.

Table 4. The share of output and highly cited papers among regional collaboration categories in SC social- and technology-oriented sub-field.

	Social-oriented SC cluster
	Output share %, 2010–2017	Output share %, 2018–2022	Top 10% cited pubs share %, 2010–2017	Top 10% cited pubs share %, 2018–2022	Top 5% cited pubs share %, 2010–2017	Top 5% cited pubs share %, 2018–2022
Single affiliation	24.52	13.20	20.94	8.12	21.4	7.09
Intra-city collaboration	43.80	40.81	36	32.94	34.16	31.25
Intranational inter-city collaboration	13.41	17.75	15.53	17.61	15.64	16.72
Intra-macro-regional collaboration	6.47	9.79	8.47	14.32	9.05	14.7
Cross-macro-regional collaboration	11.81	18.45	19.06	27.01	19.75	30.24

	Technology-oriented SC cluster
	Output share %, 2010–2017	Output share %, 2018–2022	Top 10% cited pubs share %, 2010–2017	Top 10% cited pubs share %, 2018–2022	Top 5% cited pubs share %, 2010–2017	Top 5% cited pubs share %, 2018–2022
Single affiliation	8.73	5.63	5.58	3.64	4.81	3.67
Intra-city collaboration	56.36	43.72	53.01	31.87	52.3	28.53
Intranational inter-city collaboration	13.43	17.88	11.72	17.23	11.02	17.06
Intra-macro-regional collaboration	7.58	10.04	9.38	12.2	10.02	12.29
Cross-macro-regional collaboration	13.91	22.74	20.31	35.06	21.84	38.44

As for the total of SC research, intra-city collaborations remain the primary mode of production for both fields, indicating that geographical closeness facilitates collaboration. Nevertheless, its dominance has eroded more sharply within the technology-oriented cluster, from 56.36% (2010–2017) to 43.72% (2018–2022), than in the social-oriented SC sub-field, where intra-city collaboration share has dropped slightly, from 43.8% to 40.81%.

However, when a publication comes from inter-city collaboration, there are no longer significant differences between the social and technology-focused areas. Even if a higher proportion of articles arises from collaboration between cities, countries, or macro-regions in the technology-oriented sub-field, the difference between the two sub-fields is not significant. The inter-city connectivity rises in both for the 2018–2022 period. In the case of inter-city collaborations, cross-macro-regional collaborations have the highest rates, followed by intranational inter-city links. In the case of highly cited papers, intranational inter-city collaborations are rising more and more in the technological-oriented sub-field. For the second period, the significance of cross-macro-regional collaborations increases, especially in the case of highly ‘Top 5%’ cited papers.

A critical divergence is found in the relationship between collaboration scale and citation impact. In the second period (2018–2022), the proportion of cross-macro-regional collaborations increases significantly among articles with the highest impact. In the technology-oriented sub-field there has been a dramatic pivot towards global networks. By 2018–2022, cross-macro-regional collaborations became the primary driver of elite impact, accounting for 38.44%, up from 21.84% of the ‘Top 5%’ cited papers. Simultaneously, the impact share of intra-city work in this field fell from 52.3% to 28.53%. Globalization is also significant in the social-oriented sub-field. Cross-macro-regional collaboration share rises from 19.75% to 30.24% of the ‘Top 5%’ cited publications. However, this field maintains a more balanced impact distribution, with intra-city collaboration still accounting for 31.25% of top-tier citations. These findings indicate that local context remains more vital for high-impact social-oriented research than in the technology-oriented sub-field.

5. Discussion

This study examined the global landscape of smart city research by analysing patterns of scientific production, citation impact, and inter-city collaborations. It also explored how collaboration geographies differ between technology-oriented and social-oriented research and assessed the influence of geographical distance on the resulting collaborative networks.

In line with findings from other bibliometric studies (Hoekman et al., 2010; Wuchty et al., 2007), co-authored smart city papers have a higher impact than single-author papers, and the constraining effect of geographical distance is decreasing. This trend aligns with spatial interaction theory and gravity-model mechanisms (Frenken et al., 2009b; Head & Mayer, 2014), whereby collaboration and scientific impact increase with research capacity in the city but decay with the distance between the locations. As global scientific networks expand, our results suggest that distance decay is weakening (Maisonobe et al., 2016b; Csomós et al., 2020).

Smart city research presents a spatial paradox: although more cities are participating in the global conversation, the output and impact of that research remain concentrated in defined hubs and international networks. There is a geographical split between where research is produced and where its citation impact emerges. The global North (Europe and the US) maintains the highest impact values (CNCI), even though fewer papers are produced there compared to East Asian hubs. Cities such as Sydney, Guangzhou, Nanjing, Shenzhen, and Wuhan are emerging hotspots, where both output and impact have increased. While research volume grows more slowly in European cities, their citation impact is comparatively higher. Overall, there is a shift from localized, single-institution research toward longer-distance, global collaborations, with European-Asian city partnerships playing a central role in driving high-impact research, confirming the findings by Csomós et al. (2020).

The technology-oriented and social-oriented facets of smart city research exhibit the pattern experienced between the social sciences and humanities, and sciences. Scholarly work remains biased toward technology-oriented research, which accounts for roughly two-thirds of the field, is highly collaborative, and is driven by global networks, while intra-city efforts are declining. In contrast, social-oriented research is smaller but expanding in regions such as Europe and Latin America; it tends to produce more single-author or single-institution work, and it continues to rely on local context for high-impact outcomes. These differences can be interpreted through the mechanisms highlighted in the literature: technological domains depend on infrastructure-intensive research hubs, reinforcing global concentration (Castells, 1996; Frenken et al., 2009b; Balland et al., 2020), whereas social-oriented domains are more sensitive to local context and cognitive proximity, which sustains intra-city and regional collaborations (Gertler, 2003; Boschma, 2005b).

The results indicate that geographical distance continues to influence inter-city collaborations, although its effect is declining over time. In both the social- and technology-oriented sub-fields of smart city research, intra-city collaboration remains the most common mode of production, but its share has decreased, particularly in technology-oriented research. Collaborations spanning greater distances have increased substantially – especially cross-macro-regional partnerships – and are contributing disproportionately to highly cited papers, again particularly in the technology cluster. Both sub-fields also show growth in national and macro-regional collaborations, reflecting a broader shift from localized towards more globalized scientific networks, including within social-oriented smart city research. These patterns confirm the interplay of research capacity in the cities, distance decay and disciplinary norms in shaping the spatial organization of knowledge production.

From a practical perspective, the concentration of impact in global hubs highlights opportunities and challenges for cities seeking to strengthen their smart city research capacity. Although international networks drive high-impact technological research, maintaining local embeddedness remains critical for impactful social-oriented scholarship. As cities adopt ‘smart city’ agendas (Toh, 2022), this local embeddedness will require careful urban governance approaches.

Declaration of generative AI and AI-assisted technologies in the writing process

During the preparation of this work, the authors used ChatGPT and Gemini 3 to improve language and readability. After using this tool, the authors reviewed and edited the content as needed and take full responsibility for the content of the publication.

Data availability

Underlying data

The metadata was collected from the Web of Science (WoS) Core Collection database https://www.webofscience.com. As the WoS database is provided by a third party, Clarivate, the raw data cannot be shared. The search strategy was guided by the definitions of subject categories (the SC field in WoS) and established conceptual frameworks (Cocchia, 2014; Albino et al., 2015; Eremia et al., 2017), as detailed in Table 1.

Zenodo: Data and Code for the manuscript The Global Inter-City Collaboration Network of Smart City Research: A Spatial Scientometrics Analysis (2010–2022). https://doi.org/10.5281/zenodo.19254448 (Vida & Borsi, 2026).

This project contains the following underlying data:

- SCfigure_map_code.R (analysis code used in the study).
- c1.cityid.mregio.csv (Cities’ coordinates by publications. Randomized publication ID, cities’ coordinates, country, state, city, macro region names. Cities’ coordinates source: https://simplemaps.com/data/world-cities downloaded 2023.07.)
- kw.countKL.csv (Variables for word clouds, 2010–2017. cluster ID, keywords, frequency of keywords).
- kw.countKL2.csv (Variables for word clouds, 2018–2022. cluster ID, keywords, frequency of keywords).
- kw.p.cl2.p1.csv (Variables about smart city social- and technology-oriented clusters and collaboration categories, 2010–2017. randomized publication ID, cluster ID, collaboration category ID).
- kw.p.cl2.p2.csv (Variables about smart city social- and technology-oriented clusters and collaboration categories, 2018–2022. randomized publication ID, cluster ID, collaboration category ID).
- kw.p.clus.1.csv (Variables about smart city clusters and collaboration categories, 2010–2017. randomized publication ID, cluster ID, collaboration category ID).
- kw.p.clus.2.csv (Variables about smart city clusters and collaboration categories, 2018–2022. randomized publication ID, cluster ID, collaboration category ID).
- p1.imp.csv (Variables about impact of inter-city collaboration publications, 2010–2017. randomized publication ID, standardized CNCI, Dummy for top5, Dummy for top10, for 3–5 spatial collaboration categories).
- p1.imp.all.csv (Variables about impact of publications, 2010–2017. randomized publication ID, standardized CNCI, Dummy for top5, Dummy for top10, for 1–5 spatial collaboration categories).
- p2.imp.csv (Variables about impact of inter-city collaboration publications, 2018–2022. randomized publication ID, standardized CNCI, Dummy for top5, Dummy for top10, for 3–5 spatial collaboration categories).
- p2.imp.all.csv (Variables about impact of publications, 2018–2022. randomized publication ID, standardized CNCI, Dummy for top5, Dummy for top10, for 1–5 spatial collaboration categories).

Data are available under the terms of the Creative Commons Attribution 4.0 International license (CC-BY 4.0).

References

Abramo G, D’Angelo CA, Di Costa F: Knowledge spillovers: Does the geographic proximity effect decay over time? A discipline-level analysis, accounting for cognitive proximity, with and without self-citations. J. Informet. 2020; 14(4): 101072. Publisher Full Text
Autant-Bernard C, Mairesse J, Massard N: Spatial knowledge diffusion through collaborative networks. Pap. Reg. Sci. 2007; 86(3): 341–350. Publisher Full Text
Albino V, Berardi U, Dangelico RM: Smart Cities: Definitions, Dimensions, Performance, and Initiatives. J. Urban Technol. 2015; 22(1): 3–21. Publisher Full Text
Balland P-A, Jara-Figueroa C, Petralia SG, et al.: Complex economic activities concentrate in large cities. Nat. Hum. Behav. 2020; 4(3): 248–254. PubMed Abstract | Publisher Full Text
Becher T, Trowler P: Academic tribes and territories. McGraw-Hill Education (UK); 2001.
Bibri SE, Krogstie J: The emerging data–driven Smart City and its innovative applied solutions for sustainability: the cases of London and Barcelona. Energy Inform. 2020; 3(1): 5. Publisher Full Text
Bibri SE, Krogstie J: Smart sustainable cities of the future: An extensive interdisciplinary literature review. Sustain. Cities Soc. 2017; 31: 183–212. Publisher Full Text
Blondel VD, Guillaume JL, Lambiotte R, et al.: Fast unfolding of communities in large networks. J. Stat. Mech. Theory Exp. 2008; 2008(10): P10008. Publisher Full Text
Borsi B, Soós S, Vida Z: Keyword Standardization and Restructuring: The Impact on Analysing Network-based Science Maps in Innovation Management Research. Scientometrics. 2025; 130: 593–617. Publisher Full Text
Boschma R: Editorial: Role of Proximity in Interaction and Performance: Conceptual and Empirical Challenges. Reg. Stud. 2005a; 39(1): 41–45. Publisher Full Text
Boschma R: Proximity and Innovation: A Critical Assessment. Reg. Stud. 2005b; 39(1): 61–74. Publisher Full Text
Bu J, Emmitt S, Adeyeye K: Mapping minds: geometry of cognitive distance in interdisciplinary design teams. Archit. Eng. Des. Manag. 2025; 1-21: 1–21. Publisher Full Text
Callon M, Courtial JP, Turner WA, et al.: From translations to problematic networks: An introduction to co-word analysis. Soc. Sci. Inf. 1983; 22(2): 191–235. Publisher Full Text
Cao Z, Derudder B, Dai L, et al.: An analysis of the evolution of Chinese cities in global scientific collaboration networks. Manuscript prepared for special issue on “China’s internationalization and changing role in the world”. ZFW – Advances in Economic Geography. 2023; 67(1): 5–19. Publisher Full Text
Castells M: The rise of the network society. Oxford: Blackwell Publishers Ltd; 1996.
Cocchia A: Smart and Digital City: A Systematic Literature Review. Dameri RP, Rosenthal-Sabroux C, editors.Smart City: How to Create Public and Economic Value with High Technology in Urban Space.Cham: Springer International Publishing; 2014; pp. 13–43. Publisher Full Text
Csomós G: A spatial scientometric analysis of the publication output of cities worldwide. J. Informet. 2018; 12(2): 547–566. Publisher Full Text
Csomós G, Vida ZV, Lengyel B: Exploring the changing geographical pattern of international scientific collaborations through the prism of cities. PLOS ONE. 2020; 15(11): e0242468. PubMed Abstract | Publisher Full Text | Free Full Text
Derudder B, Taylor PJ: Three Globalizations Shaping the Twenty-first Century: Understanding the New World Geography through Its Cities. Ann. Am. Assoc. Geogr. 2020; 110(6): 1831–1854. Publisher Full Text
Duranton G, Puga D: Chapter 48 - Micro-Foundations of Urban Agglomeration Economies.Henderson JV, Thisse J-F, editors. Handbook of Regional and Urban Economics. Elsevier; 2004; Vol. 4. : pp. 2063–2117. Publisher Full Text
Eremia M, Toma L, Sanduleac M: The Smart City Concept in the 21st Century. Procedia Eng. 2017; 181: 12–19. Publisher Full Text
Fernández A, Ferrándiz E, León MD: Proximity dimensions and scientific collaboration among academic institutions in Europe: The closer, the better?. Scientometrics. 2016; 106(3): 1073–1092. Publisher Full Text
Frenken K, Hardeman S, Hoekman J: Spatial scientometrics: Towards a cumulative research program. J. Informet. 2009a; 3(3): 222–232. Publisher Full Text
Frenken K, Hoekman J, Kok S, et al.: Death of Distance in Science? A Gravity Approach to Research Collaboration.Pyka A, Scharnhorst A, editors. Innovation Networks. Understanding Complex Systems. Berlin, Heidelberg: Springer; 2009b. Publisher Full Text
Frenken K, Hoekman J: Spatial Scientometrics and Scholarly Impact: A Review of Recent Studies, Tools, and Methods.Ding Y, Rousseau R, Wolfram D, editors. Measuring Scholarly Impact: Methods and Practice. Springer International Publishing; 2014; pp. 127–146. Publisher Full Text
Friedmann J: The World City Hypothesis. Dev. Chang. 1986; 17(1): 69–83. Publisher Full Text
Fujita M, Krugman P, Venables A: The spatial economy.1999; 1.
Gao S: Spatial Scientometrics.Schintler LA, McNeely CL, editors. Encyclopedia of Big Data. Springer International Publishing; 2022; pp. 872–874. Publisher Full Text
Gertler MS: Tacit knowledge and the economic geography of context, or the undefinable tacitness of being (there). J. Econ. Geogr. 2003; 3(1): 75–99. Publisher Full Text
Hall KL, Vogel AL, Huang GC, et al.: The science of team science: A review of the empirical evidence and research gaps on collaboration in science. Am. Psychol. 2018; 73(4): 532–548. PubMed Abstract | Publisher Full Text
Hamers L, Hemeryck Y, Herweyers G, et al.: Similarity measures in scientometric research: the Jaccard index versus Salton's cosine formula. Inf. Process. Manag. 1989; 25(3): 315–318. Publisher Full Text
Head K, Mayer T: Chapter 3 - Gravity Equations: Workhorse,Toolkit, and Cookbook.Gopinath G, Helpman E, Rogoff K, editors. Handbook of International Economics. Elsevier; 2014; Vol. 4. : pp. 131–195. Publisher Full Text
Hidalgo CA, Balland P-A, Boschma R, et al.: The Principle of Relatedness. Unifying Themes in Complex Systems IX. Cham: 2018. Publisher Full Text
Hoekman J, Frenken K, Tijssen RJW: Research collaboration at a distance: Changing spatial patterns of scientific collaboration within Europe. Res. Policy. 2010; 39(5): 662–673. Publisher Full Text
Kessler MM: Bibliographic coupling between scientific papers. Am. Doc. 1963; 14: 10–25. Publisher Full Text
Le Pennec E, Slowikowski K: ggwordcloud: A word cloud geom for 'ggplot2' (Version 0.6.2) [Computer software].2024. Reference Source
Maisonobe M, Eckert D, Grossetti M, et al.: The world network of scientific collaborations between cities: domestic or international dynamics?. J. Informet. 2016a; 10(4): 1025–1036. Publisher Full Text
Maisonobe M, Grossetti M, Milard B, et al.: The Global Evolution of Scientific Collaboration Networks Between Cities (1999–2014). Rev. Fr. Sociol. 2016b; 57(3): 417–441. Publisher Full Text
Massicotte P, South A: rnaturalearth: World map data from natural earth (Version 1.2.0) [Computer software].2026. Reference Source
Matthiessen CW, Schwarz AW, Find S: World Cities of Scientific Knowledge: Systems, Networks and Potential Dynamics. An Analysis Based on Bibliometric Indicators. Urban Stud. 2010; 47(9): 1879–1897. Publisher Full Text
Mora L, Deakin M, Reid A: Combining co-citation clustering and text-based analysis to reveal the main development paths of smart cities. Technol. Forecast. Soc. Chang. 2019a; 142: 56–69. Publisher Full Text
Mora L, Deakin M, Reid A: Strategic principles for smart city development: A multiple case study analysis of European best practices. Technol. Forecast. Soc. Chang. 2019b; 142: 70–97. Publisher Full Text
Pan RK, Kaski K, Fortunato S: World citation and collaboration networks: uncovering the role of geography in science. Sci. Rep. 2012; 2(1): 902. PubMed Abstract | Publisher Full Text | Free Full Text
Pebesma E: Simple features for R: Standardized support for spatial vector data. R. I. Dent. J. 2018; 10(1): 439–446. Publisher Full Text
Pebesma E, Bivand R: Spatial data science: With applications in R. Chapman and Hall/CRC; 2023. Publisher Full Text
Rasoulzadeh Aghdam S, Bababeimorad B, Ghasemzadeh B, et al.: Social smart city research: interconnections between participatory governance, data privacy, artificial intelligence and ethical sustainable development [Systematic Review]. Front. Sustain. Cities. 2025; 6–2024. Publisher Full Text
Ricciardi F, Za S: Smart City Research as an Interdisciplinary Crossroads: A Challenge for Management and Organization Studies. Mola L, Pennarola F, Za S, editors. From Information to Smart Society.Cham: Springer; 2015; pp. 149–162. (Lecture Notes in Information Systems and Organisation; vol 5. ). Publisher Full Text
Salton G, McGill MJ: Introduction to modern information retrieval. Auckland: McGraw-Hill; 1983.
Sandip T: The U.S. national nanotechnology infrastructure network and support of photonics research and development. Proceedings Volume 6182, Photonic Crystal Materials and Devices III (i.e. V). 2006; 618206. Publisher Full Text
Sassen S: The global city: New York, London, Tokyo. Princeton, NJ: Princeton University Press; 1989.
Sassen S: The Global City: Introducing a Concept. Brown J. World Aff. 2005; 11(2): 27–43. Reference Source
Shang S, Chen L, Jensen CS, et al.: Introduction to Spatio-temporal data management and analytics for Smart City research. GeoInformatica. 2020; 24(1): 1–2. Publisher Full Text
Slowikowski K: ggrepel: Automatically position non-overlapping text labels with 'ggplot2' (Version 0.9.6) [Computer software].2024. Reference Source
Small H: Co-citation in the scientific literature: A new measure of the relationship between two documents. J. Am. Soc. Inf. Sci. 1973; 24: 265–269. Publisher Full Text
South A, Michael S, Massicotte P: rnaturalearthdata: World vector map data from natural earth used in 'rnaturalearth' (Version 1.0.0) [Computer software].2024. Reference Source
Taylor PJ: Specification of the World City Network. Geogr. Anal. 2001; 33(2): 181–194. Publisher Full Text
Toh CK: Smart city indexes, criteria, indicators and rankings: An in-depth investigation and analysis. IET Smart Cities. 2022; 4(3): 211–228. Publisher Full Text
van Raan A : Measuring science: Basic principles and application of advanced bibliometrics.Glänzel W, editor. Springer handbook of science and technology indicators. Springer; 2019; pp. 237–280.
Verginer L, Riccaboni M: Talent goes to global cities: The world network of scientists’ mobility. Res. Policy. 2021; 50(1): 104127. PubMed Abstract | Publisher Full Text | Free Full Text
Vida ZV, Borsi B: Data and Code for the manuscript The Global Inter-City Collaboration Network of Smart City Research: A Spatial Scientometrics Analysis (2010–2022). [Data set]. Zenodo. 2026. Publisher Full Text
Wickham H: ggplot2: Elegant graphics for data analysis. Springer-Verlag; 2016. Reference Source
Wuchty S, Jones BF, Uzzi B: The Increasing Dominance of Teams in Production of Knowledge. Science. 2007; 316(5827): 1036–1039. PubMed Abstract | Publisher Full Text
Wuestman ML, Hoekman J, Frenken K: The geography of scientific citations. Res. Policy. 2019; 48(7): 1771–1780. Publisher Full Text
Yan E: Disciplinary knowledge production and diffusion in science. J. Assoc. Inf. Sci. Technol. 2016; 67(9): 2223–2245. Publisher Full Text
Yan E, Ding Y: Scholarly network similarities: How bibliographic coupling networks, citation networks, cocitation networks, topical networks, coauthorship networks, and coword networks relate to each other. J. Am. Soc. Inf. Sci. Technol. 2012; 63: 1313–1326. Publisher Full Text
Zhang L, Gong J, Yang Y, et al.: Cities in global and sustainable scientific collaboration network: unveiling cooperation centers. Humanit. Soc. Sci. Commun. 2025a; 12(1): 1406. Publisher Full Text
Zhang Y, He G, Jiang Z: Mapping science and revealing disciplinary communication modalities via pre-trained graph neural networks. J. Informet. 2025b; 19(4): 101741. Publisher Full Text
Zhao F, Fashola OI, Olarewaju TI, et al.: Smart city research: A holistic and state-of-the-art literature review. Cities. 2021; 119: 103406. Publisher Full Text
Zupic I, Čater T: Bibliometric methods in management and organization. Organ. Res. Methods. 2015; 18(3): 429–472. Publisher Full Text

Comments on this article Comments (0)

Version 1

VERSION 1 PUBLISHED 29 Apr 2026

Author details Author details

¹ Dept. of Science Policy & Scientometrics, Library and Information Centre of the Hungarian Academy of Sciences, Budapest, 1051, Hungary
² Dept. of Strategy, Entrepreneurship and International Business, ESSCA School of Management, Budapest, 1093, Hungary

Zsófia Viktória Vida
Roles: Conceptualization, Data Curation, Formal Analysis, Funding Acquisition, Investigation, Methodology, Project Administration, Validation, Visualization, Writing – Original Draft Preparation, Writing – Review & Editing

Balázs Borsi
Roles: Conceptualization, Data Curation, Formal Analysis, Methodology, Supervision, Validation, Visualization, Writing – Original Draft Preparation, Writing – Review & Editing

Competing interests

No competing interests were disclosed.

Grant information

This work was supported by the Hungarian Academy of Sciences [KGYNK2023-46]. I confirm that the funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript
The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.

Article Versions (1)

version 1

Published: 29 Apr 2026, 15:635

https://doi.org/10.12688/f1000research.179395.1

Copyright

© 2026 Vida ZV and Borsi B. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

Download

Export To

metrics

	Views	Downloads
F1000Research	-	-
PubMed Central Data from PMC are received and updated monthly.	-	-

Citations

0

SEE MORE DETAILS

CITE

how to cite this article

Vida ZV and Borsi B. The Global Inter-City Collaboration Network of Smart City Research: A Spatial Scientometrics Analysis (2010–2022) [version 1; peer review: 1 approved with reservations]. F1000Research 2026, 15:635 (https://doi.org/10.12688/f1000research.179395.1)

NOTE: If applicable, it is important to ensure the information in square brackets after the title is included in all citations of this article.

track

receive updates on this article

Track an article to receive email alerts on any updates to this article.

Open Peer Review

Version 1

VERSION 1

PUBLISHED 29 Apr 2026

Views

3

Reviewer Report 05 Jun 2026

Vedprakash Maralapalle, L. S. Raheja School of Architecture, Mumbai, India

Approved with Reservations

https://doi.org/10.5256/f1000research.197901.r487495

1. No formal gravity model estimation. The paper repeatedly invokes gravity models and spatial interaction theory in the literature review, but never formally applies them. The study claims gravity-like logic operates in scientific systems, where collaboration intensity increases with urban ... Continue reading

1. No formal gravity model estimation. The paper repeatedly invokes gravity models and spatial interaction theory in the literature review, but never formally applies them. The study claims gravity-like logic operates in scientific systems, where collaboration intensity increases with urban research mass and decreases with geographical distance. However, no regression model, no gravity equation, and no distance-elasticity coefficient is estimated. This is a major gap the paper describes gravity effects verbally without testing them statistically.
2. Absence of statistical significance tests. All comparisons between time periods and sub-fields (Tables 2, 3, 4) are purely descriptive percentages. No chi-square tests, t-tests, or non-parametric equivalents are applied. Statements like "there are no longer significant differences between the social and technology-focused areas" are not backed by any actual statistical test the word "significant" is used informally, which is misleading in a scientific paper.
3. City disambiguation is not explained. The paper identifies cities from author affiliations for 5,907 and 16,023 publications across the two periods. However, there is no explanation of how ambiguous city names were handled (e.g., "Cambridge" in the UK vs. USA, or "Springfield"). This is a well-known problem in spatial scientometrics that can significantly distort results.
4. CNCI as the only impact metric. Category Normalized Citation Impact is the sole measure of research quality. The paper does not consider other indicators such as h-index, CiteScore, or Altmetric scores. Using one citation-based metric limits the robustness of impact conclusions, especially since citation patterns differ between technology and social science domains, which is the very comparison the paper is making.
5. The "distance decay" claim is not rigorously demonstrated. The authors state that while intra-city collaborations between 2010–2017 were 46.36% of the 'Top 5%' most cited publications, this share had diminished in the second period to 29.49%. They interpret this as weakening distance decay. However, reduced intra-city share does not directly prove distance decay is weakening, it could also reflect a general increase in all types of collaboration. A proper distance decay analysis requires fitting a decay function (e.g., exponential) to distance-collaboration data, which is absent.
6. The "Big Science" model analogy is not substantiated. The conclusions state that the technology-oriented sub-field follows a globalized "Big Science" model, while social-oriented research relies more on local context and cognitive proximity. The "Big Science" concept has a specific meaning in science studies (large-scale, publicly-funded, infrastructure-intensive research like particle physics). Applying it loosely to IoT and smart grid research without justification weakens the conceptual contribution.
7. Cognitive proximity is discussed but not measured. The paper introduces cognitive proximity (Boschma, 2005) as a theoretical concept explaining social science collaboration patterns, but cognitive proximity is never operationalized or measured. It remains a verbal explanation that cannot be verified with the data used.
8. WoS data has known coverage bias. Web of Science under-represents journals from the Global South, non-English publications, and emerging-economy research institutions. Since the paper makes claims about the global distribution of SC research including the role of India and the Middle East this bias should be acknowledged and discussed as a major limitation, not merely implied.
9. The search profile may over-include irrelevant papers. The search profile is broad, including terms like "sustainable city", "liveable city", "compact city", and "green city". These terms overlap substantially with urban planning, sustainability, and environmental science literature that may not constitute "smart city" research in any meaningful sense. The authors do not validate their corpus against a manually checked sample to confirm thematic accuracy.
10. Keyword standardization procedure lacks reproducibility details. The Keyword Standardization and Restructuring (KSR) procedure involved systematic expert-supported manual cleaning in order to merge synonyms and restructure overlapping terms. However, the procedure is not described in enough detail to be reproduced. How many experts were involved? What was the inter-rater agreement? How many keywords were merged? Without this, the co-word analysis cannot be replicated.
11. The end year 2022 creates recency gaps. The dataset ends in 2022, meaning recent developments in smart city research particularly post-COVID smart infrastructure and the surge in AI-integrated urban systems are not captured. The paper does not acknowledge this as a limitation or discuss how the field may have evolved after 2022.
12. Table 2, 3, and 4 present raw percentages without confidence intervals. All tables show percentage shares for two periods, but no uncertainty or variability measures are given. For a paper presenting itself as a quantitative spatial scientometrics study, the absence of confidence intervals or standard errors for the key metrics is a notable weakness.
13. Figure 3 maps are too small to interpret. The global output and impact maps (Figure 3A–D) appear very small in the paper, making city-level detail unreadable. The paper would benefit greatly from zoomed regional inset maps for East Asia and Europe the two most important regions in the analysis.
14. The categorization of "Hot spots", "Black holes", etc. uses an arbitrary 25% threshold. Cities are categorised using a 25% above/below mean threshold for output and impact changes. This classification threshold is adopted from Matthiessen et al. (2010) without questioning whether 25% remains appropriate for the SC field or for the 2010–2022 time frame. The sensitivity of results to alternative threshold choices is never tested.
15. Generative AI use disclosure is incomplete. The authors state they used ChatGPT and Gemini 3 to improve language and readability. However, "Gemini 3" does not appear to be an existing publicly released model (Gemini versions include 1.0, 1.5, and 2.0). This suggests either a naming error or a lack of careful documentation of the AI tools used which is especially problematic in an era of growing concern about AI's role in academic writing.
16. Practical implications are vague. The discussion states that "as cities adopt smart city agendas, this local embeddedness will require careful urban governance approaches." This conclusion is too general to be actionable. The paper does not specify what governance actions are needed, for which types of cities, or how policymakers should use the findings. For a journal like F1000Research that values practical relevance, this is a significant shortcoming.

Is the work clearly and accurately presented and does it cite the current literature?

Partly
Is the study design appropriate and is the work technically sound?

Partly
Are sufficient details of methods and analysis provided to allow replication by others?

Partly
If applicable, is the statistical analysis and its interpretation appropriate?

Yes
Are all the source data underlying the results available to ensure full reproducibility?

Partly
Are the conclusions drawn adequately supported by the results?

Partly

Competing Interests: No competing interests were disclosed.

I confirm that I have read this submission and believe that I have an appropriate level of expertise to confirm that it is of an acceptable scientific standard, however I have significant reservations, as outlined above.

CITE

Report a concern

Respond or Comment

Comments on this article Comments (0)

Version 1

VERSION 1 PUBLISHED 29 Apr 2026

Open Peer Review

Reviewer Status

Reviewer Reports

	Invited Reviewers
	1
Version 1 29 Apr 26	read

Vedprakash Maralapalle, L. S. Raheja School of Architecture, Mumbai, India

Comments on this article

All Comments(0)

Add a comment

Sign up for content alerts

Browse by related subjects

Back to all reports

Reviewer Report

3 Views

05 Jun 2026 | for Version 1

Vedprakash Maralapalle, L. S. Raheja School of Architecture, Mumbai, India

3 Views Cite this report Responses(0)

Approved With Reservations

1. No formal gravity model estimation. The paper repeatedly invokes gravity models and spatial interaction theory in the literature review, but never formally applies them. The study claims gravity-like logic operates in scientific systems, where collaboration intensity increases with urban research mass and decreases with geographical distance. However, no regression model, no gravity equation, and no distance-elasticity coefficient is estimated. This is a major gap the paper describes gravity effects verbally without testing them statistically.
2. Absence of statistical significance tests. All comparisons between time periods and sub-fields (Tables 2, 3, 4) are purely descriptive percentages. No chi-square tests, t-tests, or non-parametric equivalents are applied. Statements like "there are no longer significant differences between the social and technology-focused areas" are not backed by any actual statistical test the word "significant" is used informally, which is misleading in a scientific paper.
3. City disambiguation is not explained. The paper identifies cities from author affiliations for 5,907 and 16,023 publications across the two periods. However, there is no explanation of how ambiguous city names were handled (e.g., "Cambridge" in the UK vs. USA, or "Springfield"). This is a well-known problem in spatial scientometrics that can significantly distort results.
4. CNCI as the only impact metric. Category Normalized Citation Impact is the sole measure of research quality. The paper does not consider other indicators such as h-index, CiteScore, or Altmetric scores. Using one citation-based metric limits the robustness of impact conclusions, especially since citation patterns differ between technology and social science domains, which is the very comparison the paper is making.
5. The "distance decay" claim is not rigorously demonstrated. The authors state that while intra-city collaborations between 2010–2017 were 46.36% of the 'Top 5%' most cited publications, this share had diminished in the second period to 29.49%. They interpret this as weakening distance decay. However, reduced intra-city share does not directly prove distance decay is weakening, it could also reflect a general increase in all types of collaboration. A proper distance decay analysis requires fitting a decay function (e.g., exponential) to distance-collaboration data, which is absent.
6. The "Big Science" model analogy is not substantiated. The conclusions state that the technology-oriented sub-field follows a globalized "Big Science" model, while social-oriented research relies more on local context and cognitive proximity. The "Big Science" concept has a specific meaning in science studies (large-scale, publicly-funded, infrastructure-intensive research like particle physics). Applying it loosely to IoT and smart grid research without justification weakens the conceptual contribution.
7. Cognitive proximity is discussed but not measured. The paper introduces cognitive proximity (Boschma, 2005) as a theoretical concept explaining social science collaboration patterns, but cognitive proximity is never operationalized or measured. It remains a verbal explanation that cannot be verified with the data used.
8. WoS data has known coverage bias. Web of Science under-represents journals from the Global South, non-English publications, and emerging-economy research institutions. Since the paper makes claims about the global distribution of SC research including the role of India and the Middle East this bias should be acknowledged and discussed as a major limitation, not merely implied.
9. The search profile may over-include irrelevant papers. The search profile is broad, including terms like "sustainable city", "liveable city", "compact city", and "green city". These terms overlap substantially with urban planning, sustainability, and environmental science literature that may not constitute "smart city" research in any meaningful sense. The authors do not validate their corpus against a manually checked sample to confirm thematic accuracy.
10. Keyword standardization procedure lacks reproducibility details. The Keyword Standardization and Restructuring (KSR) procedure involved systematic expert-supported manual cleaning in order to merge synonyms and restructure overlapping terms. However, the procedure is not described in enough detail to be reproduced. How many experts were involved? What was the inter-rater agreement? How many keywords were merged? Without this, the co-word analysis cannot be replicated.
11. The end year 2022 creates recency gaps. The dataset ends in 2022, meaning recent developments in smart city research particularly post-COVID smart infrastructure and the surge in AI-integrated urban systems are not captured. The paper does not acknowledge this as a limitation or discuss how the field may have evolved after 2022.
12. Table 2, 3, and 4 present raw percentages without confidence intervals. All tables show percentage shares for two periods, but no uncertainty or variability measures are given. For a paper presenting itself as a quantitative spatial scientometrics study, the absence of confidence intervals or standard errors for the key metrics is a notable weakness.
13. Figure 3 maps are too small to interpret. The global output and impact maps (Figure 3A–D) appear very small in the paper, making city-level detail unreadable. The paper would benefit greatly from zoomed regional inset maps for East Asia and Europe the two most important regions in the analysis.
14. The categorization of "Hot spots", "Black holes", etc. uses an arbitrary 25% threshold. Cities are categorised using a 25% above/below mean threshold for output and impact changes. This classification threshold is adopted from Matthiessen et al. (2010) without questioning whether 25% remains appropriate for the SC field or for the 2010–2022 time frame. The sensitivity of results to alternative threshold choices is never tested.
15. Generative AI use disclosure is incomplete. The authors state they used ChatGPT and Gemini 3 to improve language and readability. However, "Gemini 3" does not appear to be an existing publicly released model (Gemini versions include 1.0, 1.5, and 2.0). This suggests either a naming error or a lack of careful documentation of the AI tools used which is especially problematic in an era of growing concern about AI's role in academic writing.
16. Practical implications are vague. The discussion states that "as cities adopt smart city agendas, this local embeddedness will require careful urban governance approaches." This conclusion is too general to be actionable. The paper does not specify what governance actions are needed, for which types of cities, or how policymakers should use the findings. For a journal like F1000Research that values practical relevance, this is a significant shortcoming.

Is the work clearly and accurately presented and does it cite the current literature?

Partly
Is the study design appropriate and is the work technically sound?

Partly
Are sufficient details of methods and analysis provided to allow replication by others?

Partly
If applicable, is the statistical analysis and its interpretation appropriate?

Yes
Are all the source data underlying the results available to ensure full reproducibility?

Partly
Are the conclusions drawn adequately supported by the results?

Partly

Competing Interests

No competing interests were disclosed.

I confirm that I have read this submission and believe that I have an appropriate level of expertise to confirm that it is of an acceptable scientific standard, however I have significant reservations, as outlined above.

Respond to this report

Responses (0)

[1] Abramo G, D’Angelo CA, Di Costa F: Knowledge spillovers: Does the geographic proximity effect decay over time? A discipline-level analysis, accounting for cognitive proximity, with and without self-citations. J. Informet. 2020; 14(4): 101072. Publisher Full Text

[2] Autant-Bernard C, Mairesse J, Massard N: Spatial knowledge diffusion through collaborative networks. Pap. Reg. Sci. 2007; 86(3): 341–350. Publisher Full Text

[3] Albino V, Berardi U, Dangelico RM: Smart Cities: Definitions, Dimensions, Performance, and Initiatives. J. Urban Technol. 2015; 22(1): 3–21. Publisher Full Text

[4] Balland P-A, Jara-Figueroa C, Petralia SG, et al.: Complex economic activities concentrate in large cities. Nat. Hum. Behav. 2020; 4(3): 248–254. PubMed Abstract | Publisher Full Text

[5] Becher T, Trowler P: Academic tribes and territories. McGraw-Hill Education (UK); 2001.

[6] Bibri SE, Krogstie J: The emerging data–driven Smart City and its innovative applied solutions for sustainability: the cases of London and Barcelona. Energy Inform. 2020; 3(1): 5. Publisher Full Text

[7] Bibri SE, Krogstie J: Smart sustainable cities of the future: An extensive interdisciplinary literature review. Sustain. Cities Soc. 2017; 31: 183–212. Publisher Full Text

[8] Blondel VD, Guillaume JL, Lambiotte R, et al.: Fast unfolding of communities in large networks. J. Stat. Mech. Theory Exp. 2008; 2008(10): P10008. Publisher Full Text

[9] Borsi B, Soós S, Vida Z: Keyword Standardization and Restructuring: The Impact on Analysing Network-based Science Maps in Innovation Management Research. Scientometrics. 2025; 130: 593–617. Publisher Full Text

[10] Boschma R: Editorial: Role of Proximity in Interaction and Performance: Conceptual and Empirical Challenges. Reg. Stud. 2005a; 39(1): 41–45. Publisher Full Text

[11] Boschma R: Proximity and Innovation: A Critical Assessment. Reg. Stud. 2005b; 39(1): 61–74. Publisher Full Text

[12] Bu J, Emmitt S, Adeyeye K: Mapping minds: geometry of cognitive distance in interdisciplinary design teams. Archit. Eng. Des. Manag. 2025; 1-21: 1–21. Publisher Full Text

[13] Callon M, Courtial JP, Turner WA, et al.: From translations to problematic networks: An introduction to co-word analysis. Soc. Sci. Inf. 1983; 22(2): 191–235. Publisher Full Text

[14] Cao Z, Derudder B, Dai L, et al.: An analysis of the evolution of Chinese cities in global scientific collaboration networks. Manuscript prepared for special issue on “China’s internationalization and changing role in the world”. ZFW – Advances in Economic Geography. 2023; 67(1): 5–19. Publisher Full Text

[15] Castells M: The rise of the network society. Oxford: Blackwell Publishers Ltd; 1996.

[16] Cocchia A: Smart and Digital City: A Systematic Literature Review. Dameri RP, Rosenthal-Sabroux C, editors.Smart City: How to Create Public and Economic Value with High Technology in Urban Space.Cham: Springer International Publishing; 2014; pp. 13–43. Publisher Full Text

[17] Csomós G: A spatial scientometric analysis of the publication output of cities worldwide. J. Informet. 2018; 12(2): 547–566. Publisher Full Text

[18] Csomós G, Vida ZV, Lengyel B: Exploring the changing geographical pattern of international scientific collaborations through the prism of cities. PLOS ONE. 2020; 15(11): e0242468. PubMed Abstract | Publisher Full Text | Free Full Text

[19] Derudder B, Taylor PJ: Three Globalizations Shaping the Twenty-first Century: Understanding the New World Geography through Its Cities. Ann. Am. Assoc. Geogr. 2020; 110(6): 1831–1854. Publisher Full Text

[20] Duranton G, Puga D: Chapter 48 - Micro-Foundations of Urban Agglomeration Economies.Henderson JV, Thisse J-F, editors. Handbook of Regional and Urban Economics. Elsevier; 2004; Vol. 4. : pp. 2063–2117. Publisher Full Text

[21] Eremia M, Toma L, Sanduleac M: The Smart City Concept in the 21st Century. Procedia Eng. 2017; 181: 12–19. Publisher Full Text

[22] Fernández A, Ferrándiz E, León MD: Proximity dimensions and scientific collaboration among academic institutions in Europe: The closer, the better?. Scientometrics. 2016; 106(3): 1073–1092. Publisher Full Text

[23] Frenken K, Hardeman S, Hoekman J: Spatial scientometrics: Towards a cumulative research program. J. Informet. 2009a; 3(3): 222–232. Publisher Full Text

[24] Frenken K, Hoekman J, Kok S, et al.: Death of Distance in Science? A Gravity Approach to Research Collaboration.Pyka A, Scharnhorst A, editors. Innovation Networks. Understanding Complex Systems. Berlin, Heidelberg: Springer; 2009b. Publisher Full Text

[25] Frenken K, Hoekman J: Spatial Scientometrics and Scholarly Impact: A Review of Recent Studies, Tools, and Methods.Ding Y, Rousseau R, Wolfram D, editors. Measuring Scholarly Impact: Methods and Practice. Springer International Publishing; 2014; pp. 127–146. Publisher Full Text

[26] Friedmann J: The World City Hypothesis. Dev. Chang. 1986; 17(1): 69–83. Publisher Full Text

[27] Fujita M, Krugman P, Venables A: The spatial economy.1999; 1.

[28] Gao S: Spatial Scientometrics.Schintler LA, McNeely CL, editors. Encyclopedia of Big Data. Springer International Publishing; 2022; pp. 872–874. Publisher Full Text

[29] Gertler MS: Tacit knowledge and the economic geography of context, or the undefinable tacitness of being (there). J. Econ. Geogr. 2003; 3(1): 75–99. Publisher Full Text

[30] Hall KL, Vogel AL, Huang GC, et al.: The science of team science: A review of the empirical evidence and research gaps on collaboration in science. Am. Psychol. 2018; 73(4): 532–548. PubMed Abstract | Publisher Full Text

[31] Hamers L, Hemeryck Y, Herweyers G, et al.: Similarity measures in scientometric research: the Jaccard index versus Salton's cosine formula. Inf. Process. Manag. 1989; 25(3): 315–318. Publisher Full Text

[32] Head K, Mayer T: Chapter 3 - Gravity Equations: Workhorse,Toolkit, and Cookbook.Gopinath G, Helpman E, Rogoff K, editors. Handbook of International Economics. Elsevier; 2014; Vol. 4. : pp. 131–195. Publisher Full Text

[33] Hidalgo CA, Balland P-A, Boschma R, et al.: The Principle of Relatedness. Unifying Themes in Complex Systems IX. Cham: 2018. Publisher Full Text

[34] Hoekman J, Frenken K, Tijssen RJW: Research collaboration at a distance: Changing spatial patterns of scientific collaboration within Europe. Res. Policy. 2010; 39(5): 662–673. Publisher Full Text

[35] Kessler MM: Bibliographic coupling between scientific papers. Am. Doc. 1963; 14: 10–25. Publisher Full Text

[36] Le Pennec E, Slowikowski K: ggwordcloud: A word cloud geom for 'ggplot2' (Version 0.6.2) [Computer software].2024. Reference Source

[37] Maisonobe M, Eckert D, Grossetti M, et al.: The world network of scientific collaborations between cities: domestic or international dynamics?. J. Informet. 2016a; 10(4): 1025–1036. Publisher Full Text

[38] Maisonobe M, Grossetti M, Milard B, et al.: The Global Evolution of Scientific Collaboration Networks Between Cities (1999–2014). Rev. Fr. Sociol. 2016b; 57(3): 417–441. Publisher Full Text

[39] Massicotte P, South A: rnaturalearth: World map data from natural earth (Version 1.2.0) [Computer software].2026. Reference Source

[40] Matthiessen CW, Schwarz AW, Find S: World Cities of Scientific Knowledge: Systems, Networks and Potential Dynamics. An Analysis Based on Bibliometric Indicators. Urban Stud. 2010; 47(9): 1879–1897. Publisher Full Text

[41] Mora L, Deakin M, Reid A: Combining co-citation clustering and text-based analysis to reveal the main development paths of smart cities. Technol. Forecast. Soc. Chang. 2019a; 142: 56–69. Publisher Full Text

[42] Mora L, Deakin M, Reid A: Strategic principles for smart city development: A multiple case study analysis of European best practices. Technol. Forecast. Soc. Chang. 2019b; 142: 70–97. Publisher Full Text

[43] Pan RK, Kaski K, Fortunato S: World citation and collaboration networks: uncovering the role of geography in science. Sci. Rep. 2012; 2(1): 902. PubMed Abstract | Publisher Full Text | Free Full Text

[44] Pebesma E: Simple features for R: Standardized support for spatial vector data. R. I. Dent. J. 2018; 10(1): 439–446. Publisher Full Text

[45] Pebesma E, Bivand R: Spatial data science: With applications in R. Chapman and Hall/CRC; 2023. Publisher Full Text

[46] Rasoulzadeh Aghdam S, Bababeimorad B, Ghasemzadeh B, et al.: Social smart city research: interconnections between participatory governance, data privacy, artificial intelligence and ethical sustainable development [Systematic Review]. Front. Sustain. Cities. 2025; 6–2024. Publisher Full Text

[47] Ricciardi F, Za S: Smart City Research as an Interdisciplinary Crossroads: A Challenge for Management and Organization Studies. Mola L, Pennarola F, Za S, editors. From Information to Smart Society.Cham: Springer; 2015; pp. 149–162. (Lecture Notes in Information Systems and Organisation; vol 5. ). Publisher Full Text

[48] Salton G, McGill MJ: Introduction to modern information retrieval. Auckland: McGraw-Hill; 1983.

[49] Sandip T: The U.S. national nanotechnology infrastructure network and support of photonics research and development. Proceedings Volume 6182, Photonic Crystal Materials and Devices III (i.e. V). 2006; 618206. Publisher Full Text

[50] Sassen S: The global city: New York, London, Tokyo. Princeton, NJ: Princeton University Press; 1989.

[51] Sassen S: The Global City: Introducing a Concept. Brown J. World Aff. 2005; 11(2): 27–43. Reference Source

[52] Shang S, Chen L, Jensen CS, et al.: Introduction to Spatio-temporal data management and analytics for Smart City research. GeoInformatica. 2020; 24(1): 1–2. Publisher Full Text

[53] Slowikowski K: ggrepel: Automatically position non-overlapping text labels with 'ggplot2' (Version 0.9.6) [Computer software].2024. Reference Source

[54] Small H: Co-citation in the scientific literature: A new measure of the relationship between two documents. J. Am. Soc. Inf. Sci. 1973; 24: 265–269. Publisher Full Text

[55] South A, Michael S, Massicotte P: rnaturalearthdata: World vector map data from natural earth used in 'rnaturalearth' (Version 1.0.0) [Computer software].2024. Reference Source

[56] Taylor PJ: Specification of the World City Network. Geogr. Anal. 2001; 33(2): 181–194. Publisher Full Text

[57] Toh CK: Smart city indexes, criteria, indicators and rankings: An in-depth investigation and analysis. IET Smart Cities. 2022; 4(3): 211–228. Publisher Full Text

[58] van Raan A : Measuring science: Basic principles and application of advanced bibliometrics.Glänzel W, editor. Springer handbook of science and technology indicators. Springer; 2019; pp. 237–280.

[59] Verginer L, Riccaboni M: Talent goes to global cities: The world network of scientists’ mobility. Res. Policy. 2021; 50(1): 104127. PubMed Abstract | Publisher Full Text | Free Full Text

[60] Vida ZV, Borsi B: Data and Code for the manuscript The Global Inter-City Collaboration Network of Smart City Research: A Spatial Scientometrics Analysis (2010–2022). [Data set]. Zenodo. 2026. Publisher Full Text

[61] Wickham H: ggplot2: Elegant graphics for data analysis. Springer-Verlag; 2016. Reference Source

[62] Wuchty S, Jones BF, Uzzi B: The Increasing Dominance of Teams in Production of Knowledge. Science. 2007; 316(5827): 1036–1039. PubMed Abstract | Publisher Full Text

[63] Wuestman ML, Hoekman J, Frenken K: The geography of scientific citations. Res. Policy. 2019; 48(7): 1771–1780. Publisher Full Text

[64] Yan E: Disciplinary knowledge production and diffusion in science. J. Assoc. Inf. Sci. Technol. 2016; 67(9): 2223–2245. Publisher Full Text

[65] Yan E, Ding Y: Scholarly network similarities: How bibliographic coupling networks, citation networks, cocitation networks, topical networks, coauthorship networks, and coword networks relate to each other. J. Am. Soc. Inf. Sci. Technol. 2012; 63: 1313–1326. Publisher Full Text

[66] Zhang L, Gong J, Yang Y, et al.: Cities in global and sustainable scientific collaboration network: unveiling cooperation centers. Humanit. Soc. Sci. Commun. 2025a; 12(1): 1406. Publisher Full Text

[67] Zhang Y, He G, Jiang Z: Mapping science and revealing disciplinary communication modalities via pre-trained graph neural networks. J. Informet. 2025b; 19(4): 101741. Publisher Full Text

[68] Zhao F, Fashola OI, Olarewaju TI, et al.: Smart city research: A holistic and state-of-the-art literature review. Cities. 2021; 119: 103406. Publisher Full Text

[69] Zupic I, Čater T: Bibliometric methods in management and organization. Organ. Res. Methods. 2015; 18(3): 429–472. Publisher Full Text

The Global Inter-City Collaboration Network of Smart City Research: A Spatial Scientometrics Analysis (2010–2022)

Abstract

Background

Methods

Results

Conclusions

Keywords

1. Introduction

2. Literature review

2.1. The spatial organization of academic knowledge production

2.2. Inter-city scientific networks and the global city

2.3. Thematic heterogeneity and field-specific collaboration patterns

2.4. Smart City (SC) research: An interdisciplinary knowledge domain

3. Data and methods

3.1. Data retrieval and scope

Table 1. Smart City (SC) research search profile.

3.2. Data processing

Figure 1. Wordclouds of Keywords of Clusters in Smart City Domain.

Figure 2. Wordclouds of Keywords for social- and technology-oriented Smart City clusters

Figure 3. Output (number of publications) and impact (mean Category Normalized Citation Index, CNCI) of Smart City (SC) Research.

Figure 4. Scientific Output-Impact change between 2010–2017 to 2018–2022 in SC research.

Figure 5. Output (number of publications) and Impact (mean Category Normalized Citation Index, CNCI) of Smart City (SC) social-oriented and technology-oriented sub-fields.

Figure 6. Scientific output and impact change from 2010–2017 to 2018–2022 in the social-oriented sub-field of SC research.

Figure 7. Scientific Output-Impact change between 2010–2017 to 2018–2022 in social-oriented sub-field of SC research without the outliers.

Figure 8. Scientific Output-Impact change between 2010–2017 to 2018–2022 in the technology-oriented domain of SC research.

4. Results

4.1. The global distribution of SC research (RQ1)

Table 2. The share of intra- and inter-city collaboration based on research output and highly cited papers in SC field.

4.2. The global distribution and comparison of SC sub-fields: technology-oriented vs. social-oriented (RQ2)

4.3. Geographical distance effect on inter-city collaboration networks in SC research (RQ3)

Table 3. The share of output by regional collaboration categories in SC field.

Table 4. The share of output and highly cited papers among regional collaboration categories in SC social- and technology-oriented sub-field.

5. Discussion

Declaration of generative AI and AI-assisted technologies in the writing process

Data availability

Underlying data

References

Comments on this article Comments (0)

Open Peer Review

Comments on this article Comments (0)

Open Peer Review

Reviewer Status

Reviewer Reports

Comments on this article

Browse by related subjects

Competing Interests Policy

Stay Updated