6⊙L.MORA ET AL focusing only on the overall intellectual structure of this knowledge domain expanding the number of databases used to source documents3 including both academic publications and gray literature. This bibliometric study used 1,067 source documents identified with a keyword search and combining the analysis of the citations between them,together with citation and pub- lication counts,which are the two most basic bibliometric measures (Tijssen and van Leeuwen,2003;Martin and Daim,2008).These documents represent all the smart-city lit- erature published in the English language between 1992 and 2012.More specifically,that English language literature containing the term“smart city”or the term“smart cities,”in the title,abstract,keyword list,or body of the text,and stored in the following scholarly databases:Google Scholar;ISI Web of Science;IEEE Xplore;Scopus;SpringerLink; Engineering Village;ScienceDirect;and Taylor and Francis Online. The use of multiple databases made it possible to conduct a comprehensive interdisci- isnany plinary search and broaden the field of investigation,avoiding the risk of not capturing the full extent of research on smart cities.However,it is important to note that this choice was 9 particularly challenging and time consuming because the initial number of publications identified with the keyword search was 9,799.To extract the list of source documents, each publication was included in a single dataset and checked to correct typographical errors in the titles,authors'names,or publication dates.Repeated documents that were found in more than one database were eliminated.Finally,the title,abstract,keyword list,and body of the text of each remaining publication was manually examined to verify the effective presence of the keyword.Documents in which this search was shown to be negative were eliminated. After completing the search phase,the source documents were cataloged considering their type:abstracts,editorials,journal articles,books,book chapters,conference papers,and gray literature(See Figure 1).The last category includes the documents gen- erally defined as gray and represents a substantial part of the scientific production, especially in recent years(Schopfel and Farace,2010).According to the most common 五 definition,gray literature represents the literature that is"produced on all levels of govern- ment,academics,business and industry in print and electronic formats,but [...not con- apeoluMo trolled by commercial publishers,i.e.where publishing is not the primary activity of the producing body"(Schopfel,2010:12). All source documents were then linked to authors by their full names and the organ- izations they represent.Details about organizations were found by searching their official websites,the source documents,and the databases used for the keyword search.In this study,the most recent affiliation was attributed to each author.During this activity, data on both the type and location of each organization was also collected(See Figure 2).Based on types,four main categories were identified:(1)research and education:uni- versities,academies,and colleges;(2)research and business:private companies operating in the ICT sector which are involved in research and consultancy activities or in the dis- tribution of goods and services;(3)research and government:public authorities and their research institutes);(4)other.In case of organizations operating in multiple locations,the main headquarters were considered. Finally,before starting the analysis,citation data were extracted manually from the list of references included in each source document.In addition,considering that citation data. focusing only on the overall intellectual structure of this knowledge domain . expanding the number of databases used to source documents3 . including both academic publications and gray literature. This bibliometric study used 1,067 source documents identified with a keyword search and combining the analysis of the citations between them, together with citation and pub￾lication counts, which are the two most basic bibliometric measures (Tijssen and van Leeuwen, 2003; Martin and Daim, 2008). These documents represent all the smart-city lit￾erature published in the English language between 1992 and 2012. More specifically, that English language literature containing the term “smart city” or the term “smart cities,” in the title, abstract, keyword list, or body of the text, and stored in the following scholarly databases:4 Google Scholar; ISI Web of Science; IEEE Xplore; Scopus; SpringerLink; Engineering Village; ScienceDirect; and Taylor and Francis Online.5 The use of multiple databases made it possible to conduct a comprehensive interdisci￾plinary search and broaden the field of investigation, avoiding the risk of not capturing the full extent of research on smart cities. However, it is important to note that this choice was particularly challenging and time consuming because the initial number of publications identified with the keyword search was 9,799. To extract the list of source documents, each publication was included in a single dataset and checked to correct typographical errors in the titles, authors’ names, or publication dates. Repeated documents that were found in more than one database were eliminated. Finally, the title, abstract, keyword list, and body of the text of each remaining publication was manually examined to verify the effective presence of the keyword. Documents in which this search was shown to be negative were eliminated. After completing the search phase, the source documents were cataloged considering their type: abstracts, editorials, journal articles, books, book chapters, conference papers,6 and gray literature7 (See Figure 1). The last category includes the documents gen￾erally defined as gray and represents a substantial part of the scientific production, especially in recent years (Schopfel and Farace, 2010). According to the most common definition, gray literature represents the literature that is “produced on all levels of govern￾ment, academics, business and industry in print and electronic formats, but […] not con￾trolled by commercial publishers, i.e. where publishing is not the primary activity of the producing body” (Schopfel, 2010: 12). All source documents were then linked to authors by their full names and the organ￾izations they represent. Details about organizations were found by searching their official websites, the source documents, and the databases used for the keyword search. In this study, the most recent affiliation was attributed to each author. During this activity, data on both the type and location of each organization was also collected (See Figure 2). Based on types, four main categories were identified: (1) research and education: uni￾versities, academies, and colleges; (2) research and business: private companies operating in the ICT sector which are involved in research and consultancy activities or in the dis￾tribution of goods and services; (3) research and government: public authorities and their research institutes); (4) other. In case of organizations operating in multiple locations, the main headquarters were considered. Finally, before starting the analysis, citation data were extracted manually from the list of references included in each source document. In addition, considering that citation data 6 L. MORA ET AL. Downloaded by [Shanghai Jiaotong University] at 03:36 25 August 2017