A Bibliometric Diagnosis and Analysis about Smart Cities

This article aims to present a bibliometric analysis of Smart Cities. The study analyzes the most important journals during the period between 1991 and 2019. It provides helpful insights into the document types, the distribution of countries/territories, the distribution of institutions, the authors’ geographical distribution, the most active authors and their research interests or fields, the relationships between principal authors and more relevant publications, and the most cited articles. This paper also provides important information about the core and historical references and the most cited papers. The analysis used the keywords and thematic noun-phrases in the titles and abstracts of the sample papers to explore the hot research topics in the top journals (e.g., ‘Smart Cities’, ‘Intelligent Cities’, ‘Sustainable Cities’, ‘e-Government’, ‘Digital Transformation’, ‘Knowledge-Based City’, etc.). The main objective is to have a quantitative description of the published literature about Smart Cities; this description will be the basis for the development of a methodology for the diagnosis of the maturity of a Smart City. The results presented here help to define the scientific concept of Smart Cities and to measure the importance that the term has gained through the years. The study has allowed us to know the main indicators of the published literature in depth, from the date of publication of the first articles and the evolution of these indicators to the present day. From the main indicators in the literature, some were selected to be applied: The most influential journals on Smart Cities according to the general citation structure in Smart Cities, Global Impact Factor of Smart Cities, number of publications, publications on Smart Cities around the world, and their correlation.


Introduction
This article presents an exhaustive review of the bibliography of Smart Cities with worldwide scope and over all time. The search for and gathering of data were performed considering that there were publications previous to the appearance and use of the term "Smart City" and the acknowledgement of the scientific community.The identification of this term's appearance was made by previous studies that were devoted to investigating the evolution of concepts and uses of other terms throughout history before "Smart City" became the most used term. The concept of the Smart City was used tacitly before being recognized as a term by the disciplines, institutions, and entities that work in the creation, study, and improvement of cities.
A bibliometric study offers a statistical description of scientific production. It tries to extract as much information as possible from the studied data set to offer researchers a complete and organized vision of the multidisciplinary scientific production of the subject matter studied.
the Baltic countries of Lithuania, Latvia, and Estonia between 1996 and 2018 using bibliometric analysis. In qualitative terms (citation rate and h-index), the article ranked highest the countries with which authors from these countries collaborated, as well as the main journals and authors [31]. Furthermore, Ahmad, P. et al. [32], proposed the article "A Bibliometric Analysis of Periodontology 2000". Periodontology 2000 is a publication of 100 articles produced by eminent researchers and clinicians from many dental institutions and countries.
Bibliometrics can also be used simultaneously with other tools, as in the article "Text Mining Using Database Tomography and Bibliometrics: A Review", which describes the Database Tomography (D.T.), a textual database analysis system consisting of two major components: Algorithms for extracting multi-word phrase frequencies and phrase proximity and the interpretative capabilities of the expert human analyst, proposed by Kostoff, R.N. [33].
Some publications use bibliometrics as a part of a methodology, e.g., the article "Evaluating 'Payback' on Biomedical Research from Papers Cited in Clinical Guidelines: Applied Bibliometric Study". This article presents the development of a methodology for evaluating the impact of the research on health care (Grant, J. et al. [34]). In addition, in the article entitled "Bibliometric Analysis of Global Trends for Research Productivity in Microbiology" presents a bibliometric analysis of publications between 1995 and 2003 on microbiology (Vergidis, P.I. et al. [35]).
Some articles are focused on or limited to only one journal or magazine, and even when the focus is the same unique magazine, each author uses and combines different tools according to their needs.
For instance, articles on engineering use graphic tools; e.g., the article "Bibliometric Study of the Journal Ingeniería (2010-2017)" analyzes the complete bibliography and authors considering productivity, authorship, citation, subject, and geographic coverage, along with collaboration networks, thematic conceptual maps, and impact metrics. Ingeniería is a scientific journal edited in Colombia by Universidad Distrital Francisco Jose de Caldas. This publication reviews more than 144 papers [36].
Bibliometrics could also be used to measure changes; e.g., the article "Trends and Changes in Thunderbird International Business Review", written by Ratten, V. et al. [37]. Thunderbird International Business Review is amongst the most influential journals in the field of international business studies. Another example in this type of study is the one focused on the evolution of knowledge, e.g."Fifty Years of the Financial Review: A Bibliometric Overview" written by Baker et al. [38].
There is also the case of articles focused on several journals, such as "A Bibliometric Analysis of the Conversion and Reporting of Pilot Studies Published in Six Anaesthesia Journals" by Charlesworth, M. et al. [39], or the one written by Van Noorden, R. et al. [40] analyzing the top 100 most cited research papers of all time.
Articles proposing comparisons are also available; e.g., the article "Bibliographic and Web Citations: What Is the Difference?" by Vaughan, L. [41], which presents the differences between these concepts by comparing 46 journals in library and information science, or the article "Interdisciplinary Research by the Numbers", written by Van Noorden, R. [42], which analyzes the interactions among several disciplines (143 specialities) and their impacts in science.
There are also articles comparing authors and magazines simultaneously; e.g., "A Bibliometric Analysis of Articles Identified by Editors as Representing Excellence in Nursing Publication" analyzes subsequent citations of articles identified by editors as reflecting excellence in nursing literature and a companion dataset from the same journals comparing the concepts of reach, persistence, and dissemination in these two datasets (Nicoll, L.H. et al. [43]).
Bibliometrics is also consistently used in engineering and science. Some articles focus on particular topics; e.g., the article entitled "Comprehensive Analysis of Energy Management Strategies for Hybrid Electric Vehicles (HVes) Based on Bibliometrics" written by Zhang, P. et al. [44] quantitatively analyzes the current research status of energy management strategies of HVes.
Some articles use only some tools of bibliometrics; e.g., the article "Using Data-Sets from the Web of Science (WoS)". This study conducts a co-word analysis of 1971 publications on customer relationship management from East Asia, North America, and Europe and uses WoS as the source (written by Liu, W. et al. [45]). Another article classified in this category is the article "Bibliometrics and Beyond: Some Thoughts on Web-Based Citation Analysis", which presents in-depth research on citation analysis and the evolution from the citation index to the bibliometric spectroscopy concept (written by Cronin,B. [46]). In this category, we can include the article "Bibliometric Indicators: Quality Measurements of Scientific Publication" by Durieux, V. [47], which provides an overview of the currently used bibliometric indicators and summarizes the critical elements and characteristics that one should be aware of when evaluating the quantity and quality of scientific output. Other examples in this category are the articles proposed by Guerola-Navarro V. [48], Vicedo, P. [49], and Gil-Gómez H. [50] as studies preceding an industrial process optimization.
There are bibliometric studies for forecasting based on the evolution of publications across the years; e.g., the article "Forecasting Emerging Technologies: Use of Bibliometrics and Patent Analysis" by Daim, T.U. [51] makes forecasts for three emerging technology areas by integrating the use of bibliometrics and patent analysis into well-known technology forecasting tools, such as scenario planning, growth curves, and analogies.
Other studies use other databases but the same tools; e.g., the article entitled "The Eigenfactor algorithm and Impact Factor (IF)", published online in Journal Citation Reports as part of the ISI Web of Knowledge, which was also analyzed by Fersht, A. [52] in the article "The Most Influential Journals: Impact Factor and Eigenfactor". The analysis of other indexes with the WoS is also convenient, as in the article "Mapping of Drinking Water Research: A Bibliometric Analysis of Research Output during 1992-2011", where Fu, H. et al. [53] present a bibliometric analysis based on the Science Citation Index Expanded from the WoS. The article provides insights into research activities and tendencies of global drinking water from 1992 to 2011. The author also applied the procedure in the article "A Bibliometric Analysis of Solid Waste Research during the Period 1993-2008". The authors, Fu, H.Z. et al. [54], analyze aspects including document type, language, and publication output as well as the distribution of journals, subject category, countries, institutes, title-words, and author.
From articles mentioned in the last paragraph and many of the articles reviewed, the average period of bibliometrics studies is twenty years. For instance, the text of "Global Urbanization Research from 1991 to 2009, A Systematic Research Review", written by Wang, H. et al. [55], analyzes scientific outputs, subject categories, significant journals, international collaboration, geographic distribution, and temporal trends in keyword usage in urbanization.
Other articles also based in the Web of Science have allowed us to define the methodology; e.g., the text of the article "The Bibliometric Analysis of Scholarly Production", which is an article studying the ways that institutions and universities of science are ranked worldwide (written by Ellegaard, O. [56]). However, reviewing the literature on bibliometrics has allowed for the definition of the information to be gathered, the contents of tables, the period, and the other main aspects. However, the review of these approximately fifty articles also led to the construction of the methodology presented in the following paragraph.

Methodology
There is no previous methodology describing a standardized procedure indicating the number of papers as proper regarding the number of published or referenced papers in a specific database, the structure of matrices and tables, or specific indicators. The methodology presented here was defined from the ideas, analyses, and conclusions of previous research presented in the literature review section-the main factor considered for the definition of a statistically significant sample that guarantees representative results. The methodology was determined after reviewing data and noting that the distribution of the number of citations is concentrated in a few articles. With the support of the conclusions in these articles, we determined that the more representative variables are the number of works published, citations, the Impact Factor (IF), and the h-index. Thus, if a set of papers has an h-index of 30, it means that at least 30 papers have each received 30 citations or more. This measure combines the number of papers with citations [20].
The data were obtained from a query of the WoS database, one of the most important databases in the world, which guarantees the representativeness of the data. A series of indicators accepted to analyze the data and used by researchers of high relevance in the field of bibliometrics were used. Although the database does not include all the journals and all the articles written at all times, the most impactful journals specialized in the areas related to Smart Cities are found in the WoS.
Regarding authorship, this research aims to identify mainly productivity, identifying those authors who publish the highest numbers of papers independently, whether these papers are single-authored or not.
The primary materials used in this research are data lodged in the WoS Database and tools incorporated into the WoS system for the classification and analysis of data. The method used could be described as the organization of these data in tables to try to get valuable information that could help readers know the evolution of research on the Smart City so far through analysis of statistics.
The first step was to select the databases to be used for the recovery of the articles. Databases gathering sciences and areas of knowledge related to the theme studied and most relevant were selected: Academic Search by EBSCO Publishing, Arnetminer (Aminer) by the German Archaeological Institute and the University of Cologne, the Scopus abstract and citation database of peer-reviewed research literature, the Science Citation Index (SCI), Social Sciences Citation Index (SSCI), and Humanities Citation Index (A&HCI) of the Web of Science (WoS),the Association for Computing Machinery, and the Digital Library of IEEE Xplore. After reviewing the previous articles, functionality, and availability of databases, the Web of Science was chosen as the tool for data recovery and analysis.
The Web of Science (WoS) is a platform based on web technology that gathers the references of the prominent scientific publications of any discipline of knowledge-scientific, technological, humanistic, or sociological-essential for the support of research and the recognition of the efforts and advances made by the scientific and technical community.
The second step was the identification of the type of documents to be analyzed. There are many types of publications: Articles, meetings, books, reviews, editorials, clinical trials, corrections, letters, data papers, biographies, and retracted publications. Nevertheless, many of them are discarded for this study because only those that introduce a scientific contribution are taken into account: Articles, reviews, notes, and letters.
A scientific text is a written production that addresses theories, concepts, or any other subject based on scientific knowledge through a specialized technical language. It should be emphasized that scientific publications represent more than 50% of all publications. Table 2 presents the types of documents or formats in which the information is presented.
The search "Smart Cit*" retrieves 17,774 documents. Using filter tools and limiting the types of documents to articles, books, and reviews, the number of documents retrieved is 10,357 (58% of total records).
The analysis was focused on the results obtained from "Smart Cit*" referred to in articles, books, and reviews. The search term that was selected by its statistical frequency was Smart City.
The third step was the revision of articles on bibliometrics and selection of indicators, a combination of variables to get an appropriate analysis.
The top five journals all publish articles in all areas of technology rather than in a particular branch. Therefore, this article analyzes publications in the five journals as a dataset in this study. The sample-set is composed of the documents that were published by the five journals.
The Web of Science (WoS) database provides users access to a wide range of bibliographic and citation information from articles published in international journals over a long period.
Articles were reviewed to select the most proper database, which is to say the database with more resources, available data, and tools to facilitate the comparison and analysis according to the criteria of time, authors, and publishers. The WoS database collection indexes documents of different types, namely, articles, reviews, proceedings papers, editorial material, and book reviews, in various languages.
The fourth step was defining the period representative and useful for accomplishing the research objectives. This study collects and analyzes documents of all types that were written in English between 1991 and 2019.
The "analyze the results" tool of the Web of Science (WoS) database allows for classifications of "authors", "countries", "document types", "organizations", "publication years", and "source titles". The WoS also has the "create citation report" tool, which allows the collection of information relating to the "sum of the times cited" and "average citations per item".
The versatility of the analysis tools of the Web of Science allows filtering of the data to obtain the most detailed data possible. So, it is possible to know the organized data in the way that is needed for the analysis that we want to develop: We want to select from the search for the leading publications and know from them the number of citations and publications in the field of Smart Cities to establish a distribution of the number of publications with a minimum number of citations (at least 200, 100, or 50), as well as the h-index and the Impact Factor (IF). This procedure can be replicated, but taking institutions and countries as variables, rather than publications.
The Web of Science also allows distribution of publications to know the data from a temporal perspective, that is, classifying the number of publications each year, determining the number of publications for each one, and making a comparison with the years n − 1 and n − 2 and their respective Impact Factors.
For data analysis, Microsoft Excel (2019) was used. The tables were created and distributed in a comprehensive dashboard, facilitating analysis and contrast among them. The WoS database allows the researcher to download the data in text format, which can be transferred to tools such as Microsoft Excel (2019) and Microsoft Power BI (Pro) to create a dynamic dashboard. Through tables, graphs, and dynamic tables, the data are classified and organized to extract the main conclusions.
The Impact Factor (IF) is a measure of the frequency with which the average article in a journal has been cited in a particular year. It is used to measure the importance or rank of a journal by calculating the times its articles are cited (https://clarivate.com/webofsciencegroup/essays/impact-factor/).
The main 300 articles were ranked in terms of numbers of articles published by a specific magazine regarding Smart Cities; the magazine with the most articles published is considered as the most important and the first one.
Previously to the analysis, the data were treated to eliminate duplicity.

Duplicates
It was necessary to identify and analyze the possible duplicity of values in two specific variables: Countries and journals. When analyzing these variables, among the main countries, China (first place) and People's Republic of China (second place) were referenced; also in the case of journals, Sensors of Basel, Switzerland and Sensors as a Journal from MDPI are the same reviews. Finally, it was demonstrated that there is no duplicity when both terms are selected for refining the results, as the system shows the real number of articles.
Another case of often-recurring duplicates is the appearance of the United Kingdom and England. In addition, a productive and influential institution is found by not only the publications of its researchers, but also by the collaborations with researchers from other institutions.

Results and Discussion
This section presents the results obtained by the implementation of the methodology exposed in the previous section. The figures and tables are based on the data retrieved from WoS for the most prolific authors, institutions, and countries regarding Smart Cities. The tables and the figures are based on the aforementioned variables, both individually and as matrices, resulting in a combination of them.
The results are finally summarized in the following tables and figures: Document types ( Table 2, presented in the methodology section (Section 3)), identification of main terms (Table 1, also mentioned in Section 3), most influential journals on Smart Cities according to the WoS (Table 3), general citation structure in Smart Cities (Table 4), global impact factor of Smart Cities (Table 5), number of publications ("Smart Cit*") ( Figure 1), Smart Cities publications around the world (Figure 2), the most productive and influential institutions (Table 6), the most productive countries in Smart Cities (Table 7), the most productive and influential authors (Table 8), authors with the highest numbers of papers in the top four journals (Table 9), institutions with the highest numbers of papers in the top four journals (Table 10), the most productive countries and journals in Smart Cities (Table 11), and the 300 most cited papers on Smart Cities of all time (Table A1) .
As shown in the Table 4, only four articles have been cited more than 500 times in all time, and two between 2011 and 2019. These articles, presented in Table 4, have all been published in the last decade, the oldest of them dating from 2009, and the most recent of 2019 is the most cited of all time.
As shown in Table 5, the scientific production of the decade between 2010 and 2019 is 20 times higher.
The years 2016 and 2017 are the years with the highest Impact Factors and the most productivity.
In terms of the first factor considered in the bibliometric studies, as seen in Figure 1, for the geographical distribution of the scientific production, the United States (1567 articles) and China (1224 articles) stand out as the most relevant countries. Nevertheless, if the European Union is considered a whole, it stands out as the geographical area with the highest scientific production in the field of Smart Cities (6640 articles), followed by the United States and Canada, which have produced more than 2500 articles on this subject. On the other hand, Russia, one of the most advanced countries, is one of those lagging behind among the most relevant countries regarding Smart Cities. Figure 1 presents the distribution of the contributions of countries to the debate around Smart Cities, This figure shows the most prolific European countries in the field of Smart Cities. Italian researchers have published more than 800 articles, followed by Spain with more than 700 articles published, and by researchers in the United Kingdom with more than 650 articles. After Europe and the United States, the largest scientific production is located in Asia. Countries like India, China, and Japan have produced more than 1200 articles.    Between 2015 and 2018, there was a great leap, and more than a thousand articles were published. In 2017, 1652 articles were published, which represents 614 articles more than the previous year.   Table 6 presents the most productive institutions in terms of Smart Cities. It summarizes the Total Papers (TP) and the Total Citations (TC) in journals indexed in WoS; >200, >100, and >50 = the number of papers with more than 200, 100, and 50 citations. It also summarizes the number of papers, their citations in the last ten years (P10Y and C10Y, respectively), and the Impact Factor (IF). Note that only one institution from the United States is among the most relevant, while five institutions from Italy and four from China are there. China is the country with the most citations in this table, while England and Italy are essential from the perspective of most cited institutions. Table 7 presents the most productive countries in terms of Smart Cities. It summarizes TP and TC (total papers and citations in journals indexed in WoS, respectively), the numbers of papers with more than 200, 100, and 50 citations (>200, >100, >50), P10Y and C10Y (number of papers and their citations in the last ten years), and the Impact Factor (IF).
Considering the results as a whole, there is a strong correlation between academia, industrial development, and the strengthening of Smart Cities.
The number of articles and citations is low compared with other topics; however, the results in this research demonstrate that Smart Cities are becoming a transcendental subject in the current scenario of societies. New countries and institutions are starting to participate in this global discussion of the digitalization of urban centers. Furthermore, the number of authors and media involved in Smart Cities research and dissemination is increasing across the years. Wang, Y.

Conclusions
The analysis shows that Smart City research is a theme where many areas of research converge. The bibliometric analysis indicates that Smart Cities are emerging as a fast-growing topic of scientific inquiry, and much of the knowledge generated about them is singularly technological. A Smart City is a social and economic phenomenon driven by environmental issues and human welfare.
A specific methodology was defined to take advantage of data available in the Web of Science (WoS). The types of documents were also selected to obtain a bibliographic study, including only those written from a scientific perspective. In the matter of Smart Cities, some publications do not have the scientific character expected for this study; then, they were excluded from the study. Table 2 shows the papers' distribution by the type of document, which was obtained using the "analyze the results" tool of the WoS database. Overall, 46% of the papers were articles, 44% were meeting papers, and 4% were books. The proportion of other types (news, letters, corrections, reviews, biographical items) was small (less than 5%).
As presented in Table 3, the most influential journals on Smart Cities, according to the WoS, are Sensors (Basel-Open Access Journal), IEEE Access (The Multidisciplinary Open Access Journal), and Sustainability. These journals represent about 12% of the total publications about Smart Cities. In addition, IEEE Communications Magazine might be considered as an influential review; eighteen articles published in this review are referenced in the list of three hundred more relevant articles listed in Table A1 in Appendix A.
As seen in Table 4, entitled "General citation structure in Smart Cities", only four articles were cited by over five hundred citations. Furthermore, more than 96% of papers of all time were cited less than fifty times. For papers published between 2001 and 2019, the percentage of papers is similar to the percentage of papers ever published; nevertheless, those with less than fifty citations were 95%.
Many Smart Cities have evolved over the past decade. Consequently, scientific output has increased proportionally, and vice-versa. The analysis of successful cases will lead to facilitation and acceleration of the emergence of new Smart Cities. As seen in Figure 2, entitled "Annual Number of publications ('Smart Cit*')", during 2017, almost three thousand articles were published. A decade earlier, in 2007, less than 100 articles had been published, which means that the number has increased by more than thirty times. In general, the increase in the number of citations is slow; most of the articles with the majority of citations were published between 2011 and 2012.
As mentioned before, the evolution of Smart Cities is linked to technological progress, and the bibliographic study shows a high degree of correlation between the countries with the greatest technological advancement and scientific production in the field of Smart Cities. The results regarding geo-economical evolution support the still-valid influence of the United States, but show the importance of other emerging powers in terms of economy and knowledge, research, and innovation. In terms of the Smart City publications around the world, Asia is the most productive region in the world. Almost 4100 publications were published between 2012 and 2019 in Asia, while 3469 were published in Europe and 1781 in America.
The influence of educational and research institutions and universities has influenced the design, forecasting, and measurement of performance of Smart Cities. According to the results summarized in Table 6, the most prolific institution published more than twenty of the total publications in 2012; the most prolific were the Chinese Academy of Sciences, the University of London, the Consiglio Nationale delle Ricerche, the Polytechnic University of Milan, the Polytechnic University of Turin, the Massachusetts Institute of Technology, the University of Bologna, the University of Naples Federico II, the Delft University of Technology, the Wuhan University, the Centre National de la Recherche Scientifique, the Tsinghua University, the Royal Institute of Technology, the Universitat Politècnica de València, the University of Genoa, Shanghai Jiao Tong University, and The University of California Berkeley.
The leading institutions do not coincide with the most relevant countries; among the first 24 traces of the list, only three American institutions appear. This means that the United States' publications are distributed among more institutions. In contrast, in the leading countries of Table 6, China and Italy, there are institutions specialized in the topic of Smart Cities.
The most productive countries in terms of scientific publications are the United States, China, Spain, and England. It can be concluded from Table 6 ("The most productive and influential institutions"), a matrix in which the number of publications in the first thirty countries is presented, that in the journals with the largest number of publications, a large contribution is from the USA's institutions, followed by the Chinese, Spanish, Italian, and English, all of them with many publications in these magazines.
The h-index allowed us to determine the most productive authors. Table 6 presents the h-index, and this indicator is superior for the University of Bologna (26), CNR (Consiglio Nationale delle Ricerche; 22), University of London (19), and the Royal Institute of Technology (18). Table 5 presents the Global Impact Factor for Smart Cities. Between 2014 and 2017, the number of citations reached around 30,000 in 2018 and 2019. Regarding Global Impact Factor for Smart Cities, in 2012, the indicator reached a maximum value, and almost a decade later, the indicator decreased due to the number of articles published, that means, more than 13 times more.
The most productive and influential institutions do not coincide with the most relevant countries; among the first 24 traces of the list, only three American institutions appear. This means that the United States' publications are distributed among more institutions. In contrast, in the leading countries of Table 6, China and Italy, there are institutions specialized in the topic of Smart Cities.
In general, there is a very diverse relationship between authors and the journals in which they publish; there is no marked collaboration between the principal authors and the prominent journals to highlight. According to Table 9, the most prolific authors did not publish their articles in the most critical reviews in the area of Smart Cities. Nevertheless, there is a correlation between the researcher's citizenship and publications. Regarding Table 11, researchers in the USA published mainly in IEEE Access, IEEE Internet of Things, and IEEE Communications Magazine.
Sensors is the magazine with the most articles in the area of Smart Cities. This confirms the importance of these devices in Smart City development. More than 1500 articles were published and referenced in WoS. Researchers were related mainly to the USA, China, India, and South Korea. China is the country with the most productive authors (Table 8).
There is a high correlation between leading institutions and main journals ( Table 10). The three hundred most cited papers on Smart Cities are summarized in the appendix. There are 173 different magazines referenced in this list. The magazines with the most articles published were the IEEE Communications Magazine (18 articles), IEEE Access (nine articles), IEEE Internet of Things Journal (eight articles), Cities (eight articles), and Renewable and Sustainable Energy Reviews (seven articles).
Author Contributions: These authors contributed equally to this work. All authors have read and agreed to the published version of the manuscript.
Funding: This research received no external funding.

Conflicts of Interest:
The authors declare no conflict of interest.

Abbreviations
The following abbreviations are used in this manuscript: