UDC: 711.5:001.891.32:004
DOI: 10.5379/urbani-izziv-en-2021-32-01-005
Received: 5 Mar. 2021
Accepted: 28 Apr. 2021
Marjan HOCEVAR
Tomaž BARTOL
Cities as places and topics of studies: 
Mapping research clusters across disciplines
In research articles, cities usually occur as topics (e.g., sub­jects or actors) or places of studies (e.g., sites, destinations, locations, or spaces). Investigation of more general pat­terns is rare because research usually focuses on individual cities. We use science mapping, based on Scopus data and Vosviewer visualization software, to examine city-related research across journals and disciplines (subject areas), and to assess how multiple city functions are reflected in journals. Comparable European Union capital cities (Berlin, Madrid, Rome, and Warsaw) serve as models. The patterns are remarkably similar regardless of the city. Na­tional and regional journals are the most common pub­lication venues. Research takes place within three major disciplinary clusters: 1) the social sciences, and arts and humanities, 2) medicine, and 3) natural/technical scienc­es (environmental, earth and planetary, agricultural, and biological sciences). Medicine shows an early prevalence, and recently the social sciences have been strongly rep­resented in these studies. Although the relationships are based on different journals, they are comparable for all cities and can be used to assess cities of similar size. This study was conducted just before the Covid-19 pandemic, and it can serve as a reference to identify research patterns before and after because the outbreak may bring about changes in future city-related research.
Keywords: cities, science mapping, research fields, bibli­ometrics, visualization
1	Introduction
Cities feature prominently in research articles. They usually appear in different contexts: first, as an investigative field (e.g., culture, geography, history, medicine, or urban planning) and, second, specific cities as places or typologically defined areas (e.g. capital cities, Asian cities, or smart cities) and also meta­phorically (e.g., the city as a lab, network, assemblage, business, classroom, or platform). Therefore, cities can serve as topics for research (e.g., as actors, relations, brands, case studies, exam­ples, engines, experiments, factors, drivers, instruments, etc.), or come about as sites where a particular study has taken place: as a destination, location, locale, setting, site, space, and so on. Our purpose is not to theorize about how cities are studied academically.
In the context of the conceptual background of our empirical study, we merely state that the growing complexity of social, spatial, environmental, and technological intertwining is re­flected in both the heterogeneity of analyses and attempts to integrate them. This is especially true for urban and regional studies in the treatment on particular, general, and relation­al attributes of the urban (Hocevar, 2005; Cox & Evenhuis, 2020). The issue of object/subject/topic/treatment can be blurred. Thus, for example, very heterogeneous “living” actors of the city such as planners, politicians, artists, entrepreneurs, researchers, media, and so on can be treated together with “in­animate” aspects such as land, museums, manufacturing plants, publications, and studies. Both subjects and objects of the city or in the city, regardless of their affiliation, have an agency of a sort to form a network of functions, connections, or as­semblage, which can be translated into the actor as a whole. To provide a rough illustration of such an interweaving, we cite research perspectives derived from the epistemological premise of ANT (Actor–Network Theory) and socio-material assemblage concepts analysing relations between entities and their constituent elements (Latour, 2005; Brenner et al., 2011; Gutzmer, 2016). Human and nonhuman actors, so-called act­ants, function together, which is reflected in external relations. 
On the other hand, many empirically oriented articles tend to focus on individual cities where specific issues are addressed: cultural activities, the economy, the environment, food and nu­trition, health and disease, history, pollution, tourism, traffic, urban forestry, urban morphology, and so on. Although many articles focus on specific or typologically grouped cities, the work that deals with the city as a research site is less frequent or at least ambiguous when used. For example, the term labo­ratory has been used to refer to the city as a place of research and also to illustrate a type of field research (Karvonen & Heur, 2014). In addition, even research paradigms can be named after cities, such as the Chicago School of (Urban) Sociology (Gieryn, 2006; Guggenheim, 2012). In the context of ANT, the city as a laboratory figures as an actant, whether it is a metaphor or a concrete and actual site. More specifically, smart cities and urban sustainability have also been investigated as topics through quantitative bibliometric methods (Ingwers­en & Serrano-López, 2018; Wang et al., 2019; Marvuglia et al., 2020), although not based on specific cities. Kadi (2019) included all European capitals and focused on gentrification. Berlin, Rome, and Madrid (the subjects of our study) were addressed in the context of history by Therborn (2002) and Gómez et al. (2018), in the framework of city branding (de Rosa et al., 2019), and in a study on the share of publications of the world’s major agglomerations (Grossetti et al., 2014). Research fields in relation to cities were identified by Nunes et al. (2019), who used the journal classification system of re­search areas of the Web of Science.
Cities as central themes are most often found in the social sciences, where comparative studies have developed a number of quantitative and qualitative comparative techniques (Ward, 2010), which sometimes also try to balance the many divides (e.g., wealth, geography, and political systems) in comparative research (Robinson, 2011). Cities are also the subject of re­search in the life sciences; for example, in the plant sciences and environmental science (e.g., the case of Berlin; Sukopp, 2008), or climate research (Lamb et al., 2019). In this respect, the areas of urban research as identified by Raynor (2019) were more inclusive (also involving environmental factors and natural resources) but they were based only on cities in Aus­tralia. More comprehensive comparisons of different cities are however rare. In the studies pertaining to information sciences (e.g., bibliometrics and science mapping), cities are most fre­quently addressed in the sense of tracking affiliation (author’s address), as well as metropolitan units, regions, countries, and so on (Bartol & Hocevar, 2005; Frenken et al., 2009; Mat­thiessen et al., 2010). Maisonobe et al. (2017) investigated global cities in the context of scientific output and scientific disciplines. Bornmann & de Moya-Anegón (2019) examined German cities with regard to the concentration of scientific activities. Also evaluated were scientific cooperation between cities and institutions (Leydesdorff & Persson, 2010) and the output of cities on specific topics, for example, urban globali­zation (Kanai et al., 2018). Various contexts of cities based on The Rise of the Network Society (Castells, 1996) have also been “scientometrically” evaluated (Zhen et al., 2020).
Mapping of specific topics is often carried out using visuali­zation software (e.g., Vosviewer, CitNetExplorer, CiteSpace, and Pajek). Hajduk (2017) investigated city logistics. Visual­izations identified clusters of different schools of thought in the relationships between cities (Peris et al., 2018). Cities as primary sites of knowledge were also assessed for intellectual property output (i.e., patents; Kogler et al., 2018). In a study on medicine and public health, cities were evaluated on issues relevant to the aging of the population (de Oliveira et al., 2019; Xiang et al., 2020). Various bibliographic elements in publica­tions (countries, journals, title words, author keywords, etc.) were visualized in environmental sciences (e.g., cities’ ecologi­cal infrastructure; Sun et al., 2020). Terms dealing with trees and parks in the city were mapped into clusters by Xing and Brimblecombe (2020). Sometimes, the principal publications were assessed, for example, on the topic of creative cities (Rod­rigues & Franco, 2020). Article keywords (urban, city, or cities) were also used (Kirby, 2012). However, big-data approaches are also possible, taking into account the need for informed interpretation (Zook et al., 2019).
We examine and compare cities in the context of the func­tions as reflected in different areas of publishing. An addition­al motivation was the onset of the health emergency caused by severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2). However, we do not address Covid-related citations (as yet) in relation to urban matters. The time lapsed (less than one year at the time of this study) is still too short. We would rather find out what the line of research was before the crisis in order to provide a basis for possible more complex comparative studies in the future. Many previous quantitative studies that have looked at cities have focused primarily on the city as the site of an author’s institution, but this is not very informative for a city in the context of its functions. Our aim, in contrast, is to examine a city as the setting of a study or as a direct research topic, or both. Thus, we do not study a city as a bibliographic “address” but the city as an “actor”. We compare equivalent cities in different countries in order to assess parallels that perhaps transcend assumed differences. The hypothesis is that, although the cities considered in this study belong to very different geographical and linguistic con­texts and have undergone historical development, the research fronts are quite similar. We assume that this is reflected not only in the choice of academic publication channels and in a similar ratio between national and international journals, but also in the topics investigated by researchers. The hypothesis is examined through the analysis of bibliographic and text data.
The bibliographic data (i.e., growth patterns over several dec­ades, journal titles, languages, co-authorship, and country of af­filiation) provide the basis for comparison whereby elementary patterns of similarity can be assessed on the basis of weighing publishers’ data. However, these should be complemented by an assessment of distribution along disciplinary lines. Here, our preliminary pilot analysis suggests that there is a shift in emphasis toward the social sciences as well as the arts and humanities that is similar across the cities studied.
Disciplinary development can be more precisely determined by using advanced visualization software to construct maps based on text data. We not only identify distinctive clusters of re­search topics, and possible interconnections and links, but also detect the development in time that underpins preliminary information from the bibliographic data. Again, these patterns are very similar across cities, regardless of the potentially very different publication venues (e.g., journals).
2	Materials and method
Our preliminary pilot revealed an important share of city-re­lated articles in national and regional journals (also in national languages). Therefore, we chose the Scopus database over Web of Science (WOS) although WOS has recently introduced the ESCI (Emerging Sources Citation Index) in its Core Collec­tion promoting regional and specialty area publishers. Visu­alization and clustering used Vosviewer software. The analysis covers all articles up to 2019 (Scopus command: pubyear < 2020). Analysis of text data was conducted on the abstracts of the articles. The title field cannot be used because this field contains both the translated title (English) as well as the original title, where the words from the original titles skew visualizations.
The target was selected European cities. Given the very differ­ent city size, it only made sense to compare not only equivalent cities but also larger cities, which have usually more connec­tions (Levinson, 2012). In addition, a sufficient number of records are required for the visualizations to reveal applicable clusters. We checked major cities with a population around one million and upward. We excluded London, Paris, and Moscow because these cities are much larger and would thus merit a comparison on their own. Next, there was the challenge of disambiguation: does a name really refer to the city in question or perhaps only denote a phenomenon, procedure, or concept named after the city? For example, there are at least fifteen cities called Berlin or Rome (Pouliquen et al., 2006). The uni­form solution was to include both the name of the city and the respective country (both as a noun and adjective) in the retrieval. It is possible that an article is about the city, but the abstract (or keywords) will not include the country name. This is a limitation. However, the procedure of including countries is consistent because the location by country maximizes search precision (Overell & Rüger, 2008) due to the absence of am­biguities (Volz et al., 2007).
We compared, on the same principles, major European Union (capital) cities: Amsterdam (constitutional capital), Athens, Berlin, Brussels, Bucharest, Budapest, Copenhagen, Madrid, Prague, Rome, Stockholm, Vienna, and Warsaw. Accordingly, we coupled each city with the respective country noun (n.) and country adjective (a.). Two terms were usually enough, with some exceptions (Holland/Dutch/Netherlands, Denmark/Danish/Dane, etc.). Finally, we selected all (capital) cities that returned more than 2,000 articles (ar) or review articles (re), with the respective city in intersection with the corresponding country. The search statement was thus designed as follows:
(TITLE(city) AND TITLE-ABS-KEY(country-n. OR coun­try-a.) AND DOCTYPE(ar OR re) AND PUBYEAR < 2020
The cities meeting the criteria were Berlin, Madrid, Rome, and Warsaw (Table 1). We do not provide the demographic sta­tistics on these cities because these are available in many data sources (e.g., in Internet 1). According to different urban ty­pologies (cities as local administrative units, functional urban areas, metropolitan regions, etc.), Berlin, Madrid, and Rome are fairly comparable regarding their size and demographics (depending on typology). Warsaw is somewhat smaller. These cities were also ranked very similarly in studies by Csomós (2017) and Hanna and Rowley (2019).
Table 1 also provides the number of all city names (column 2: Total), in addition to the numbers applicable in this study (column 4). For example, there are 11,747 and 7,249 articles with Berlin and Rome in the title, respectively; for Madrid and Warsaw, this share is lower. Berlin, for example, frequently has only figurative historical and political connotations (e.g., Ber­lin Wall). Rome also carries references to antiquity. Thus, the city name alone is not suitable for retrieval because the names have special uses, such as metonymy or polysemy.
We also examined major non-capital cities. Munich and Ham­burg returned just over one thousand articles each, and Mi­lan about 1,270. The endonym Milano retrieved additional records. It seems that the endonym has also acquired interna­tional use. This is not the case with Rome (Ital. Roma) because the term Roma, in relation to Italy, invariably refers to the ethnic group. The major metropolis Barcelona, which features prominently in research publications, returned almost 2,500 articles. Here we also included Catalonia/Catalan. However, this city was not included in our study because we limited our study to one city per country. In this respect, Madrid was somewhat more in line with the other three cities, which are also capital cities.
3	Results and discussion
3.1	Yearly growth of articles
The increase in targeted articles has been steady, with some fluctuations (Figure 1). Before 1960, such articles are rare. Ar­ticles about Madrid show the most vigorous growth. Warsaw and Berlin present similar trends at first. The steady beginning, and then some decline (Warsaw) or surge (Berlin), only sta­bilizes (with an upward trend) in the early 1990s, likely on account of  “normalization” of political situation in central and eastern Europe.
The once-divided Berlin reflects contributions from both Ger­man states: the Federal Republic of Germany (West Germany) and the German Democratic Republic (East Germany). The place of publication cannot be ascertained very precisely giv­en the weak and deficient Scopus inclusion of country data (discussed later in this study). As inferred from journal titles (Table 2), the high counts in mid 1980s are not attributable to special events because most are mapped to the subject area of medicine. Similar applies to Warsaw. The still weak stand­ings of Warsaw in the last decade likely reflect the economic situation and lower research and development funds available to research institutions in transition (Odrobina, 2016).
3.2	Journals and articles
Subsequent records cover all Scopus articles up to 2019 (pub­year < 2020). Journals show strong “national” patterns. Most of the top five journals have an original title in the national language, although many also publish articles in English. In addition, most are mapped to medicine (Table 2). The journal names in Table 2 are spelled out as in Scopus. All consecutive words are lowercased for consistency.
Figure 2 shows journals that each published at least four ar­ticles on the subject of our study. The five principal journals (from Table 2) are located at the beginning (left side) of the x-axis. The number of articles per journal quickly decreases. The very long tail of journals that published three, two, and one article is not shown. Such inversely proportional patterns are very similar for all cities and exhibit clear characteristics of power laws. Incidentally, such power law patterns were also found in other processes contingent on the scaling functions of city size (Bettencourt et al., 2007).
We also checked some citation patterns. The journals in the same linguistic group cite similar journals quite strongly. Other studies have also detected effects of geography and distance on knowledge flows (Pan et al., 2012; Abramo et al., 2020). In our case, we assume this to be attributable to the subject addressed: the city featured in the article title and was thus relevant in a specific geographic context. However, has the representation of journals been steady throughout the years?
3.3	Country and city of affiliation, language of articles
Roughly half of all articles were (co)authored from the country of the respective city (Table 3). Moreover, between 80 and 90% of articles from these countries were co-authored by authors from that particular city. However, US and UK authors hold second and third place everywhere. Similar US/UK domi­nance was also reported in some other studies (Okorie et al., 2014). Although the most important share of all articles was published in national journals, recent attention has shifted toward international publishing, although national journals still hold top place. Shares of national languages have been decreasing, English having by now taken the principal role.
Not all articles are supplied with author-country information. The numbers for Poland should have been much higher. For example, many of the 1,206 Polish-language articles were pub­lished from an “undefined” country. Such limitations have also been reported in WOS (Liu et al., 2018), but this seems to be much more considerable in Scopus, especially for older documents. The omission of a county is much more critical than omission of language (Jacsó, 2009). We also detected deficiencies with city data. The affiliation-city field failed to correctly detect the city. Therefore, we needed to employ the complete affiliation field where, for example, Rome was fre­quently spelled out only in Italian. Warsaw was spelled out in Polish, but in various adjective/noun cases.
Co-authorship links are not very strong (Figure 3). Weak co-authorship between countries in Europe was also detected by Jokic et al. (2019). However, the US and UK are similarly linked with all countries. Visualization is thus in line with the numbers in Table 3. Articles from Spain are, on average, more recent (marked with a darker shade), and from Poland older (marked with a lighter shade), which is also in line with the data in Figure 1. In the case of Germany, we needed to construct a thesaurus in order conflate different variants. In the database country field, there were also ddr (Germ. Deutsche Demokratische Republik), frg (Federal Republic of Germany) and west ger (West Germany).
3.4	Subject areas
Figure 1 presents yearly progress, but we wished to see how this was reflected across the different research fields. Namely, research disciplines possess diverse publication patterns, and therefore journals are mapped to different subject areas. The journals in this study were mapped to twenty-seven Scopus subject areas. The patterns of scatter are remarkably similar, the top five categories being the same for all cities. Medicine is predominant, and the social sciences are ranked second. However, counts of research areas must not be generalized too much because one journal can be mapped to several ar­eas or to other areas than what one might expect (Hocevar & Bartol, 2016). Figure 4 (based on yearly growth) reveals detailed patterns: medicine has been overtaken by the social sciences for all cities. The category arts and humanities has also been progressing. Here one also needs to take into account the specifics of publication in “basic” disciplines and “strategic” disciplines (van Rijnsoever & Hessels, 2011). National research and development policies can also play a role in differences among different disciplines (Cugmas et al., 2019).
3.5	Clusters and time scale of related research topics
The last and the principal part of this study is based on text data (terms in abstracts) tackling the content of research taking place in the four cities. The terms (individual words and noun phrases) are arranged into clusters according to relatedness as detected by the software (we excluded terms from structured abstracts as well as the term city, which is found in virtually every abstract).
The first figure for each city (Figures 5, 7, 9, and 11) pre­sents clusters of related and interconnected terms indicating research areas and topics. The circles representing associated terms are marked by a distinctive shade for each cluster. The second figure (Figures 6, 8, 10, and 12; the time scale of re­search topics) shows the same terms, but focuses on develop­ment in time. The darker (the more saturated) the circle, the more recent the average year. Circle size indicates the rele­vance of a term. The strength of links within each cluster and between clusters is shown with lines. These terms provide a general idea of research emphases and approaches. Each map consists of tens of thousands of terms, and so only selected labels are visualized (to avoid overlap), usually terms with at least ten occurrences.
Employing the same principles for all maps facilitates compar­ison on the same basis. The number of terms varies among the cities, depending on the number of articles as well as city-spe­cific terminology. Identification of clusters, the timescale of av­erages, and the number of terms are defined by the algorithms explained in the program manual (van Eck & Waltman, 2019). The interpretation of clusters (i.e., research subjects) employs the categorization scheme of Scopus subject areas. All maps are based on articles from the complete Scopus database up to 2019 (pubyear < 2020).
3.5.1	Berlin
The visualization program identified 47,000 terms, which oc­cur at least ten times in 3,778 articles (Figures 5 and 6). Only the most relevant terms are shown. Three different clusters can be noticed. The strongest cluster (on the right) pertains to medicine and related research. As shown in Figure 4 (previous subsection), medicine was especially strong in earlier periods, hence some strong accents of light shades (Figure 6). This fig­ure complements Figure 5 in terms of time tendencies. The more general terms occur more frequently and are thus larger. Topics of more recent importance can be noted to the right (darker shades). Because they are recent, they occur in smaller numbers (hence smaller circles).
The lower left cluster (Figures 5 and 6) presents topics that pertain strongly to the social sciences, as well as arts and hu­manities, business, economics, and other related areas. Even though these are separate subject areas in Scopus, it is appar­ent that the research is interlinked, given the position in the cluster. Considering the recent advance of the social sciences, this cluster is expected to grow. The recent accents (darker shades) are obvious, and they include the terms activist, debate, urban space, and urban development. The most recent terms in this context are small and overlapping, and thus not visible (e.g., gentrification). Links between different clusters also exist. For example, at the bottom centre of Figures 5 and 6 are the terms medical education and public health service. This is an area of convergence between the medical and social sciences (i.e., interdisciplinarity).
The upper (smaller) cluster is defined by environmental sci­ence, followed by Earth and planetary sciences, engineering, agricultural and biological sciences, and so on. Some frequent terms can be highlighted: agriculture, contamination, species, and water.
Some contexts in which Berlin is highlighted specifically “as something”, as used verbatim in article titles, are the following: Berlin as: a case study, an example, a capital, a model, a … resort, a location, a destination, the virtual centre, the site, a creative field, a relic border, a natural and socioeconomic system, and so on. Such contexts usually pertain to the social sciences and to arts and humanities, and they can also be figurative. In other (more frequent) contexts, Berlin appears as a “matter-of-fact” place of research: “during summer in Berlin,” “micro-geograph­ic analysis for Berlin,” and so on.
3.5.2	Madrid
The maps are based on 2,954 articles and 47,000 terms (very similar to Berlin). The program identified five clusters (Fig­ure  7) with two distinguishable (interlinked) clusters related to Medicine (right side of the map). The lower medical clus­ter involves epidemiology, microbiology (disease, infection …). The upper medical cluster is more diverse, involving age, man, woman … and also blood pressure, diet, food. Just as at Berlin, the educational issues in Medicine show connections with the Social Sciences (bottom centre). Social Sciences (lower left) are again more recent and connected with Arts and Humanities. The two upper left clusters are marked by Earth and Planetary Sciences, Environmental Science, and also Agricultural and Bi­ological Sciences and Engineering, in various interconnections. The left-most cluster is located closer to the Social Sciences below and contains, on average, the oldest topics (light shades, farthest left, Figure 8). We can note a dense “sub-cluster” of geological topics. Related environmental-biological topics are more recent. The upper cluster is linked with environment and pollution. It is also connected with the upper medical cluster reflecting the impact on the health of city dwellers.
In the articles, Madrid appears as a case study, an exam­ple, a key factor in sustainable mobility, a liberation from, 
a communication ecosystem, a mercantile and financing institu­tion, and a place of research, as indicated by article titles such as “Women living homeless … Madrid”, “Epidemiology and etiol­ogy … Madrid”, and “3D groundwater flow … Madrid aquifer”.
3.5.3	 Rome
Rome returned 2,411 articles. The figures are based on 38,500 terms mapped to three different clusters (Figure 9), showing similar structure as the clusters at Berlin. Again, medicine was strong in earlier periods (right cluster, lighter shades, Figure 10). The lower left cluster (social sciences and arts and human­ities) is again more recent, showing convergence with medi­cine in the area of education. Also at Rome, the upper-left cluster is strongly linked with Earth and planetary sciences and environmental science, with terms such conservation and excavation located closer to the social sciences and the arts and humanities. At the top, there are topics related to climate and pollution, which lean toward medicine.
In the articles, Rome appears as a window of observation, a full urban experience, a stage, the source, “Spanish Avignon”, a capital, a memorial landscape, a destination, and a place of research, as indicated by article titles such as “Recreational drugs ... Rome”, “... Smart homes network in Rome”, “Com­bined magnetic, chemical ... anthropic context in Rome”, and “Urban regeneration process ... Rome”.
3.5.4	Warsaw
Warsaw is linked to 2,071 articles. The 24,000 relevant terms are mapped to four clusters (Figure 11). Medical topics are contained in two (interconnected) clusters. The lower-right cluster reflects, for example, epidemiology (with the terms dis­ease and infection). The two clusters to the left are somewhat interconnected. Many terms revolve around the generic term area with links to both clusters. The lower left cluster contains more recent research fronts (darker shades, Figure 12), which can again be labelled as related to the social sciences (and to an extent to the arts and humanities). The upper-left (smaller) cluster shows connections with Earth and planetary sciences, environmental science, and agricultural and biological scienc­es. These topics are, on average, older (lighter shades) than topics in the lower left. Average medicine-related topics are also discussed early on the time scale. This corroborates our previous assumption that the cumulative data in Table 3 no longer reflect the current situation of research articles.
In the articles found in Scopus, Warsaw appears as a commuting centre, an air transport node, an example, a chief town, an en­demic focus, and a place of research, as indicated in titles such as “Local institutions of culture as urban … Warsaw”, “Short-term impact of PM2.5, PM10 … Warsaw”, “Postmodern architecture under socialism … church in … Warsaw”.
3.6 	The four cities in perspective
The aim was not to evaluate and compare cities on the basis of an author’s affiliation, but to examine cities as objects (topics) and sites (locales) of research in order to map the multiple functions performed by heterogeneous agencies (actors) of a city as reflected in scholarly journals. 
The study was based on the evaluation of the role of four capital cities (Berlin, Madrid, Rome, and Warsaw) that have experienced some differences in the more recent historical con­text. These cities represent specific geographical and linguistic settings, ranging from the western/southern Mediterranean to central/eastern Europe. Nonetheless, the topics covered in the articles are remarkably similar across the four cities, which was our hypothesis about the similarities in disciplinary develop­ment. The clusters of topics clearly show not only comparable accents, but also similar research trends over time. The trends show a growing interest in city-related social issues.
The analysis included not only text data (disciplinary devel­opment) but also bibliographic data (scattering across publi­cations). National and regional journals are important for all four cities. One might dismiss such journals as less important in the context of high-ranking international journals, but such journals can still provide an important avenue for presenting findings based on specific cities. This can complement research published in international studies that focus on general sum­maries, possibly with multiple authorship, in which individual contributions and local subject contexts are less clear. Sim­ilarities are also reflected in co-authorship and country of affiliation. The majority of articles are published by authors from the respective city. However, international collaboration is strongest with the authors from the UK, US, and France. Again, these patterns are very similar across all four cities.
4	Conclusion
The study found that city-centred research is being conducted in a few fairly well-defined major disciplinary clusters. Where­as medicine, which used to dominate, has stagnated, the social sciences now show the strongest presence. The arts and human­ities have also made considerable progress. Less dynamic but steady is the growth of subjects in environmental science, and Earth and planetary sciences. Although journals have always played an important role in life sciences and related disciplines, it seems that journal articles have by now also been fully em­braced by the social sciences. The results of different disciplines cannot be directly compared with each other because there are many factors that influence publication behaviour in specific disciplines.
Because the publication patterns are so similar for all four cities studied, our tentative guess is that this could be true for many other cities as well. In this respect, different cities of different sizes could also be compared, which remains to be investigated.
 What will the future accents and research priorities in cit­ies be? This study assessed the state of research at the end of 2019, just before the health emergency caused by coronavirus 2 (SARS-CoV-2) or the Covid-19 pandemic. In many cases, cit­ies were particularly affected, although not all in the same way. This is where regional and national journals can play an im­portant role. The connections and research links may generate different and more innovative trends in interdisciplinary co­operation, perhaps between the social sciences and medicine, and also involving other disciplines. True interdisciplinarity has been weak to date, but it could create new relationships. This needs to be pursued further.
Marjan Hocevar, University of Ljubljana, Faculty of Social Sciences, Centre for Spatial Sociology, Ljubljana, Slovenia
E-mail: marjan.hocevar@fdv.uni-lj.si
Tomaž Bartol, University of Ljubljana, Biotechnical Faculty, Agrono­my Department, Ljubljana, Slovenia
E-mail: tomaz.bartol@bf.uni-lj.si
References
Abramo, G., D’Angelo, C. A. & Costa, F. (2020) Does the geographic proximity effect on knowledge spillovers vary across research fields? Scientometrics, 123(2), pp. 1021–1036. DOI: 10.1007/s11192-020-03411-x
Bartol, T. & Hocevar, M. (2005) The capital cities of the ten new Europe­an Union countries in selected bibliographic databases. Scientometrics, 65(2), pp. 173–187. DOI: 10.1007/s11192-005-0266-3
Bettencourt, L. M. A., Lobo, J., Helbing, D., Kühnert, C. & West, G. B. (2007) Growth, innovation, scaling, and the pace of life in cities. Pro­ceedings of the National Academy of Sciences, 104(17), pp. 7301–7306. DOI: 10.1073/pnas.0610172104
Bornmann, L. & de Moya-Anegón, F. (2019) Spatial bibliometrics on the city level. Journal of Information Science, 45(3), pp. 416–425. DOI: 10.1177/0165551518806119
Brenner, N., Madden, D. J. & Wachsmuth, D. (2011) Assemblage urban­ism and the challenges of critical urban theory. City, 15(2), pp. 225–240. DOI: 10.1080/13604813.2011.568717
Castells, M. (1996) The rise of the network society. Malden, MA, Black­well Publishers Ltd.
Cox, K. R. & Evenhuis, E. (2020) Theorising in urban and regional stud­ies: Negotiating generalisation and particularity. Cambridge Journal of Regions, Economy and Society, 13(3), pp. 425–442. 
DOI: 10.1093/cjres/rsaa036
Csomós, G. (2017) A spatial scientometric analysis of the publication output of cities worldwide. Journal of Informetrics, 11(4), pp. 976–988. DOI: 10.1016/j.joi.2017.08.006
Cugmas, M., Ferligoj, A. & Kronegger, L. (2019) Scientific co-authorship networks. In: Doreian, P., Batagelj, V. & Ferligoj, A. (eds.) Advances in network clustering and blockmodeling, pp. 363–387. Hoboken, NJ, John Wiley & Sons. DOI: 10.1002/9781119483298.ch13
de Oliveira, S. M. L., Pessa, S. L. R., Schenatto, F. J. & de Lourdes Ber­nartt, M. (2019) Cities and population aging: A literature review. In: Bagnara, S., Tartaglia, R., Albolino, S., Alexander, T. & Fujita Y. (eds.) Pro­ceedings of the 20th Congress of the International Ergonomics Association (IEA 2018), pp. 1395–1404. London, Springer International Publishing. DOI: 10.1007/978-3-319-96071-5_141
de Rosa, A. S., Bocci, E. & Dryjanska, L. (2019) Social representations of the European capitals and destination e-branding via multi-channel web communication. Journal of Destination Marketing & Management, 11, pp. 150–165. DOI: 10.1016/j.jdmm.2017.05.004
Frenken, K., Hardeman, S. & Hoekman, J. (2009) Spatial scientometrics: Towards a cumulative research program. Journal of Informetrics, 3(3), pp. 222–232. DOI: 10.1016/j.joi.2009.03.005
Gieryn, T. F. (2006) City as truth-spot: Laboratories and field-sites in urban studies. Social Studies of Science, 36(1), pp. 5–38. DOI: 10.1177/0306312705054526
Gómez, M., Fernández, A. C., Molina, A. & Aranda, E. (2018) City brand­ing in European capitals: An analysis from the visitor perspective. Jour­nal of Destination Marketing and Management, 7, pp. 190–201. Scopus. DOI: 10.1016/j.jdmm.2016.11.001
Grossetti, M., Eckert, D., Gingras, Y., Jégou, L., Larivière, V. & Milard, B. (2014) Cities and the geographical deconcentration of scientific activity: A multilevel analysis of publications (1987–2007). Urban Studies, 51(10), pp. 2219–2234. DOI: 10.1177/0042098013506047
Guggenheim, M. (2012) Laboratizing and de-laboratizing the world: Changing sociological concepts for places of knowledge production. History of the Human Sciences, 25(1), pp. 99–118. DOI:10.1177/0952695111422978
Gutzmer, A. (2016) Rethinking the city: Actor–network theory and the creation of urban connections. In: Gutzmer, A. (ed.) Urban innovation networks: Understanding the city as a strategic resource (pp. 11–21). London, Springer International Publishing. 
DOI: 10.1007/978-3-319-24624-6_3
Hajduk, S. (2017) Bibliometric analysis of publications on city logistics in international scientific literature. Procedia Engineering, 182, pp. 282–290. DOI: 10.1016/j.proeng.2017.03.194
Hanna, S. & Rowley, J. (2019) The projected destination brand personal­ities of European capital cities and their positioning. Journal of Market­ing Management, 35(11–12), pp. 1135–1158. 
DOI: 10.1080/0267257X.2019.1647274
Hocevar, M. (2005) The concept of networking of the towns—glo­balization. Sociologija i prostor: casopis za istraživanje prostornoga i sociokulturnog razvoja, 43(3), pp. 691–724. Available at: https://hrcak.srce.hr/index.php?show=clanak&id_clanak_jezik=52913 (accessed 12 Jan. 2021).
Hocevar, M. & Bartol, T. (2016) Agriculture vs. social sciences: Subject classification and sociological conceptualization of rural tourism in Scopus and Web of Science. Acta Agriculturae Slovenica, 108(1), pp. 33–44. DOI: 10.14720/aas.2016.108.1.1
Ingwersen, P. & Serrano-López, A. E. (2018) Smart city research 1990–2016. Scientometrics, 117(2), pp. 1205–1236. 
DOI: 10.1007/s11192-018-2901-9
Internet 1: https://ec.europa.eu/eurostat/cache/RCI/#?vis=city.statis­tics&lang=en (accessed 15 Apr. 2021).
Jacsó, P. (2009) Errors of omission and their implications for computing scientometric measures in evaluating the publishing productivity and impact of countries. Online Information Review, 33(2), pp. 376–385. DOI: 10.1108/14684520910951276
Jokic, M., Mervar, A. & Mateljan, S. (2019) The development of political science in central and eastern Europe: Bibliometric perspective, 1996–2013. European Political Science, 18(3), pp. 491–509. 
DOI: 10.1057/s41304-018-0191-6
Kadi, J. (2019) Which cities are studied? Probing the geographical scope of 40 years of gentrification research. Der öffentliche Sektor – The Public Sector, 45(1), pp. 49–54. DOI: 10.34749/oes.2019.3067
Kanai, J. M., Grant, R. & Jianu, R. (2018) Cities on and off the map: A bibliometric assessment of urban globalisation research. Urban Studies, 55(12), pp. 2569–2585. DOI: 10.1177/0042098017720385
Karvonen, A. & Heur, B. van (2014) Urban laboratories: Experiments in reworking cities. International Journal of Urban and Regional Research, 38(2), pp. 379–392. DOI: 10.1111/1468-2427.12075
Kirby, A. (2012) Current research on cities and its contribution to urban studies. Cities, 29(Suppl. 1), pp. S3–S8. DOI: 10.1016/j.cities.2011.12.004
Kogler, D. F., Heimeriks, G. & Leydesdorff, L. (2018) Patent portfolio analysis of cities: Statistics and maps of technological inventiveness. European Planning Studies, 26(11), pp. 2256–2278. DOI: 10.1080/09654313.2018.1530147
Lamb, W. F., Creutzig, F., Callaghan, M. W. & Minx, J. C. (2019) Learn­ing about urban climate solutions from case studies. Nature Climate Change, 9(4), pp. 279–287. DOI: 10.1038/s41558-019-0440-x
Latour, B. (2005) Reassembling the social: An introduction to actor–net­work-theory. Oxford, Oxford University Press.
Levinson, D. (2012) Network structure and city size. PLOS ONE, 7(1), e29721. DOI: 10.1371/journal.pone.0029721
Leydesdorff, L. & Persson, O. (2010) Mapping the geography of science: Distribution patterns and networks of relations among cities and insti­tutes. Journal of the American Society for Information Science and Tech­nology, 61(8), pp. 1622–1634. DOI: 10.1002/asi.21347
Liu, W., Hu, G. & Tang, L. (2018) Missing author address information in Web of Science—An explorative study. Journal of Informetrics, 12(3), pp. 985–997. DOI: 10.1016/j.joi.2018.07.008
Maisonobe, M., Grossetti, M., Milard, B., Jégou, L. & Eckert, D. (2017) The global geography of scientific visibility: A deconcentration process (1999–2011). Scientometrics, 113(1), pp. 479–493. 
DOI: 10.1007/s11192-017-2463-2
Marvuglia, A., Havinga, L., Heidrich, O., Fonseca, J., Gaitani, N. & Reck­ien, D. (2020) Advances and challenges in assessing urban sustainabili­ty: An advanced bibliometric review. Renewable and Sustainable Energy Reviews, 124, 109788. DOI: 10.1016/j.rser.2020.109788
Matthiessen, C. W., Schwarz, A. W. & Find, S. (2010) World cities of scien­tific knowledge: Systems, networks and potential dynamics. An analysis based on bibliometric indicators. Urban Studies, 47(9), pp. 1879–1897. DOI: 10.1177/0042098010372683
Nunes, D. M., Tomé, A. & Pinheiro, M. D. (2019) Urban-centric resilience in search of theoretical stabilisation? A phased thematic and concep­tual review. Journal of Environmental Management, 230, pp. 282–292. DOI: 10.1016/j.jenvman.2018.09.078
Odrobina, A. (2016) Structural barriers to research and development activities in emerging markets: The case of Poland, the Czech Republic, Slovakia and Hungary. Economics and Business Review, 16(2), pp. 39–53. DOI: 10.18559/ebr.2016.2.3
Okorie, P. N., Bockarie, M. J., Molyneux, D. H. & Kelly-Hope, L. A. (2014) Neglected tropical diseases: A systematic evaluation of research capaci­ty in Nigeria. PLoS Neglected Tropical Diseases, 8(8), e3078. 
DOI: 10.1371/journal.pntd.0003078
Overell, S. & Rüger, S. (2008) Using co-occurrence models for place­name disambiguation. International Journal of Geographical Information Science, 22(3), pp. 265–287. DOI: 10.1080/13658810701626236
Pan, R. K., Kaski, K. & Fortunato, S. (2012) World citation and collabora­tion networks: Uncovering the role of geography in science. Scientific Reports, 2(1), 902. DOI: 10.1038/srep00902
Peris, A., Meijers, E. & van Ham, M. (2018) The evolution of the systems of cities literature since 1995: Schools of thought and their interaction. Networks and Spatial Economics, 18(3), pp. 533–554. 
DOI: 10.1007/s11067-018-9410-5
Pouliquen, B., Kimler, M., Steinberger, R., Ignat, C., Oellinger, T., Blackler, K., et al. (2006) Geocoding multilingual texts: Recognition, disambigua­tion and visualisation. ArXiv:Cs/0609065. Available at: http://arxiv.org/abs/cs/0609065 (accessed 13 Feb. 2021).
Raynor, K. (2019) Australian urban scholarship as a community of prac­tice: Strengths, omissions and opportunities. Urban Policy and Research, 37(3), pp. 405–421. DOI: 10.1080/08111146.2019.1610867
Robinson, J. (2011) Cities in a world of cities: The comparative gesture. International Journal of Urban and Regional Research, 35(1), pp. 1–23. DOI: 10.1111/j.1468-2427.2010.00982.x
Rodrigues, M. & Franco, M. (2020) Networks and performance of crea­tive cities: A bibliometric analysis. City, Culture and Society, 20, 100326. DOI: 10.1016/j.ccs.2019.100326
Sukopp H. (2008) The city as a subject for ecological research. In: Marzluff, J. M. et al. (eds.) Urban ecology, pp. 281–298. Boston, Springer. DOI: 10.1007/978-0-387-73412-5_17
Sun, S., Jiang, Y. & Zheng, S. (2020) Research on ecological infrastruc­ture from 1990 to 2018: A bibliometric analysis. Sustainability, 12(6), 2304. DOI: 10.3390/su12062304
Therborn, G. (2002) Monumental Europe: The national years. On the iconography of European capital cities. Housing, Theory and Society, 19(1), pp. 26–47. DOI: 10.1080/140360902317417976
van Eck, N. J. & Waltman, L. (2019) VOSviewer manual (version 1 .6.10). Leiden, Universiteit Leiden.
van Rijnsoever, F. J. & Hessels, L. K. (2011) Factors associated with dis­ciplinary and interdisciplinary research collaboration. Research Policy, 40(3), pp. 463–472. DOI: 10.1016/j.respol.2010.11.001
Volz, R., Kleb, J. & Mueller, W. (2007) Towards ontology-based disam­biguation of geographical identifiers. Paper presented at the WWW2007: 16th International World Wide Web Conference), 8–12 May, Banff, Canada. Typescript.
Wang, M.-H., Ho, Y.-S. & Fu, H.-Z. (2019) Global performance and devel­opment on sustainable city based on natural science and social science research: A bibliometric analysis. Science of The Total Environment, 666, pp. 1245–1254. DOI: 10.1016/j.scitotenv.2019.02.139
Ward, K. (2010) Towards a relational comparative approach to the study of cities. Progress in Human Geography, 34(4), pp. 471–487. DOI: 10.1177/0309132509350239
Xiang, L., Shen, G. Q. P., Tan, Y. & Liu, X. (2020) Emerging evolution trends of studies on age-friendly cities and communities: A scientomet­ric review. Ageing & Society. DOI: 10.1017/S0144686X20000562
Xing, Y. & Brimblecombe, P. (2020) Trees and parks as “the lungs of cities.” Urban Forestry & Urban Greening, 48, 126552. 
DOI: 10.1016/j.ufug.2019.126552
Zhen, F., Tang, J. & Wang, X. (2020) How does Castells’s The Rise of the Network Society contribute to research in human geography? A cita­tion content and context analysis. The Professional Geographer, 72(1), pp. 96–108. DOI: 10.1080/00330124.2019.1611459
Zook, M., Shelton, T. & Poorthuis, A. (2019) Big data and the city. In: Schwanen, T. & van Kempen, R. (eds.) Handbook of urban geography, pp. 63–75. Cheltenham, UK, Edvard Elgar Publishing. 
DOI: 10.4337/9781785364600.00013

M. HOCEVAR, T. BARTOL

Cities as places and topics of studies: Mapping research clusters across disciplines

M. HOCEVAR, T. BARTOL

Table 1: City name occurring in the title in intersection with the respective country, and the number of records.
TITLE(city)
 Total
 TITLE-ABS-KEY(country-n. OR country-a.)
 TITLE(city) AND TITLE-ABS-KEY(country-n. OR coun­try-a.)
 
Berlin
 11,747
 Germany OR German
 3,778
 
Madrid
 4,868
 Spain OR Spanish OR Spaniard
 2,954
 
Rome
 7,249
 Italy OR Italian
 2,411
 
Warsaw
 3,065
 Poland OR Polish OR Pole
 2,071
 




Figure 1: Yearly growth of articles referring to the cities (illustration: authors).

Cities as places and topics of studies: Mapping research clusters across disciplines

Table 2: Number of journal articles by top five journals.
City and journals
 Articles
 
Berlin
 
Zeitschrift fur arztliche Fortbildung
 67
 
Gesundheitswesen
 38
 
Berliner und Munchener tierarztliche Wochenschrift
 36
 
Deutsche Gesundheitswesen
 35
 
Geburtshilfe und Frauenheilkunde
 28
 
Madrid
 
Estudios geograficos
 68
 
Revista espanola de salud publica
 67
 
Boletin geologico y minero
 42
 
Enfermedades infecciosas y microbiologia clinica
 39
 
Revista clinica espanola
 38
 
Rome
 
Nuovi annali d'igiene e microbiologia
 55
 
Annali di igiene medicina preventiva e di comunita
 53
 
Annali dell istituto superiore di sanita
 33
 
Medicina nei secoli
 31
 
Atmospheric environment
 23
 
Warsaw
 
Przeglad epidemiologiczny
 113
 
Roczniki panstwowego zakladu higieny
 101
 
Archiwum historii medycyny
 64
 
Polski tygodnik lekarski
 63
 
Przeglad lekarski
 53
 




Figure 2: Inversely proportional decreasing number of different jour­nals and articles per journal (illustration: authors).

M. HOCEVAR, T. BARTOL

Table 3: Four cities as sites or topics in article titles, authors’ affiliation, language of articles.
Articles
 Articles
 Articles
 Articles
 
Site/Topic
 Berlin
 3,778
 Madrid
 2,954
 Rome
 2,411
 Warsaw
 2,071
 
Affiliation
 Germany
 1,817
 Spain
 2,239
 Italy
 1,315
 Poland
 972
 
Berlin
 1,479
 Madrid
 1,943
 Rome/Roma
 1,190
 Warsaw/Warsz*
 907
 
US
 271
 US
 160
 US
 185
 US
 72
 
UK
 186
 UK
 98
 UK
 137
 UK
 40
 
France
 55
 France
 60
 France
 65
 Germany
 38
 
Canada
 51
 Germany
 38
 Germany
 50
 France
 15
 
Netherlands
 49
 Italy
 38
 Spain
 38
 Italy
 9
 
Switzerland
 47
 Portugal
 23
 Australia
 26
 Netherlands
 9
 
Austria
 40
 Canada
 17
 Netherlands
 21
 Belgium
 8
 
Italy
 34
 Mexico
 16
 Canada
 19
 Russian F.
 8
 
Australia
 32
 Australia
 15
 Belgium
 17
 Czech R.
 6
 
Language
 Ger
 1,772
 Spa
 1326
 Ita
 609
 Pol
 1,206
 
Eng
 1,876
 Eng
 1,678
 Eng
 1,648
 Eng
 799
 
Eng in 1995
 0.41%
 Eng in 1995
 0.43%
 Eng in 1995
 0.71%
 Eng in 1995
 0.59%
 
Eng in 2019
 0.79%
 Eng in 2019
 0.69%
 Eng in 2019
 0.94%
 Eng in 2019
 0.86%
 



Cities as places and topics of studies: Mapping research clusters across disciplines


Figure 3: Co-authorship in articles by country of authors’ affiliation (illustration: authors).


Figure 4: Growth of articles per Scopus subject area (illustration: authors).

M. HOCEVAR, T. BARTOL

Cities as places and topics of studies: Mapping research clusters across disciplines


Figure 5: Clusters of interrelated research areas and topics (Berlin) (illustration: authors). 


Figure 6: Time scale of research topics (Berlin) (illustration: authors).

M. HOCEVAR, T. BARTOL


Figure 7: Clusters of interrelated research areas and topics (Madrid) (illustration: authors).


Figure 8: Time scale of research topics (Madrid) (illustration: authors).

Cities as places and topics of studies: Mapping research clusters across disciplines


Figure 9: Clusters of interrelated research areas and topics (Rome) (illustration: authors).


Figure 10: Time scale of research topics (Rome) (illustration: authors).

M. HOCEVAR, T. BARTOL


Figure 11: Clusters of interrelated research areas and topics (Warsaw) (illustration: authors).


Figure 12: Time scale of research topics (Warsaw) (illustration: authors).

Cities as places and topics of studies: Mapping research clusters across disciplines

M. HOCEVAR, T. BARTOL

Cities as places and topics of studies: Mapping research clusters across disciplines