Visualizing Sustainability Research in Business and Management (1990–2019) and Emerging Topics: A Large-Scale Bibliometric Analysis

While researchers from many disciplines are increasingly interested in studying issues related to sustainability, few studies have presented a holistic view of sustainability from the perspectives of business and management. This bibliometric study quantitatively analyzed a big data set of 30 years of sustainability research (1990–2019), consisting of 37,322 publications and 1,199,398 cited references, visualizing major topics, dynamic evolution, and emerging development. The decade-by-decade in-depth analysis shows a clear shift from a nearly exclusive focus on economic growth and consumption to all three pillars of sustainability, i.e., economic growth, social development, and environmental protection. Highlighting the differences between United Nations’ Sustainable Development Goals and the popular research topics from academia, our analysis uncovers research gaps and suggests future research directions for sustainability researchers and practitioners.


Introduction
With the diffusions of internet and other technologies, we all have witnessed transformational development in commerce over the last thirty years. While major technology-driven companies such as Alibaba Group and Amazon have built gigantic ecosystems, many enterprises have been struggling with intense competition and shrinking market share. What leads to such a great difference? Jack Ma, the founder of Alibaba, who is also the first person to be appointed as an advocate of the United Nations' Sustainable Development Goals, provided an opinion in response to this question. According to Jack Ma (2013) [1], "Our challenge is to help more people to make healthy money, 'sustainable money,' money that is not only good for themselves but also good for the society. Business people like me are beginning to pay attention to social issues including the environment, taking action and really treating this issue very seriously." Such a sustainability-driven perspective has become increasingly popular among business researchers and practitioners. However, few studies have presented a holistic view of sustainability from the perspectives of business and management. Therefore, a bibliometric analysis of the literature on sustainability can shed fresh light on the implications of sustainability for modern commerce.  Our study makes several contributions. First, it adopts a quantitative, "big data" approach to analyze a large-scale dataset containing 37,322 sustainability-related research articles in 433 top business and management journals, and 1,199,398 cited documents collected from the Web of Science (WoS) Core Collection in the past 30 years (from 1990 to 2019). Compared with existing reviews on sustainability and related topics, our study is among the most comprehensive bibliometric reviews so far. The large dataset allowed us to capture the intellectual landscape of sustainability research and provide a comprehensive portrayal of the knowledge structure of the sustainability domain in the last three decades. Second, this review offers a detailed decade-by-decade analysis to demonstrate how the sustainability research has moved from a rather narrow, economic growth-focused agenda to encompassing all three pillars of sustainability (i.e., economic growth, social development, and environmental protection), and how it has progressed with a rich and diverse set of topics. Third, this paper suggests that business and management fields could offer more solutions to sustainability problems around the world. Business and management research has the potential to create interesting and significant topics that can link the hotspots in the sustainability literature to the United Nations' Sustainable Development Goals (SDGs). Finally, the sustainability-related studies in Sustainability journal are also mapped from 2008 to 2019 to view the extending and potential boundary of the update studies in the journal. Comparing these topics from this journal with the above studied business journals can also provide a broader vision for the journal readers.
The remaining parts of this article are organized as follows. Section 2 summarizes the related review studies. Then, the data and methods are presented in Section 3. The bibliometric analyses and results are listed in Section 4. Based on the results, the research gaps and insights are described in Section 5 and the conclusion is attained in Section 6.

Systematic Literature Reviews of Sustainability
Recognizing the responsibilities business schools have in conducting research and educating future business leaders, many scholars have reviewed different aspects of sustainability and sustainable development. For example, strategic management scholars have examined the research Our study makes several contributions. First, it adopts a quantitative, "big data" approach to analyze a large-scale dataset containing 37,322 sustainability-related research articles in 433 top business and management journals, and 1,199,398 cited documents collected from the Web of Science (WoS) Core Collection in the past 30 years (from 1990 to 2019). Compared with existing reviews on sustainability and related topics, our study is among the most comprehensive bibliometric reviews so far. The large dataset allowed us to capture the intellectual landscape of sustainability research and provide a comprehensive portrayal of the knowledge structure of the sustainability domain in the last three decades. Second, this review offers a detailed decade-by-decade analysis to demonstrate how the sustainability research has moved from a rather narrow, economic growth-focused agenda to encompassing all three pillars of sustainability (i.e., economic growth, social development, and environmental protection), and how it has progressed with a rich and diverse set of topics. Third, this paper suggests that business and management fields could offer more solutions to sustainability problems around the world. Business and management research has the potential to create interesting and significant topics that can link the hotspots in the sustainability literature to the United Nations' Sustainable Development Goals (SDGs). Finally, the sustainability-related studies in Sustainability journal are also mapped from 2008 to 2019 to view the extending and potential boundary of the update studies in the journal. Comparing these topics from this journal with the above studied business journals can also provide a broader vision for the journal readers.
The remaining parts of this article are organized as follows. Section 2 summarizes the related review studies. Then, the data and methods are presented in Section 3. The bibliometric analyses and results are listed in Section 4. Based on the results, the research gaps and insights are described in Section 5 and the conclusion is attained in Section 6.

Systematic Literature Reviews of Sustainability
Recognizing the responsibilities business schools have in conducting research and educating future business leaders, many scholars have reviewed different aspects of sustainability and sustainable Sustainability 2019, 11, 5596 4 of 37 development. For example, strategic management scholars have examined the research in corporate social responsibility [10]; in behavioral economics, economists have examined if behavioral nudges can lead to green choices [11]; marketing scholars have reviewed the findings on eco-friendly product development [12] and ethical consumption behaviors [13]; operations management focuses on supply chain management [14], smart technologies in manufacturing [15], supply chain collaboration [16], circular economy [17], sustainable transportation systems [18], and sustainability performance evaluation [19]; the human resources discipline focuses on the role of human resources management [20]; and entrepreneurship scholars have examined topics such as social entrepreneurship [21]. Some reviews are interdisciplinary due to the nature of the sustainability-related issues investigated, such as resilience [22], internet use and individual well-being [23], and social and sustainable innovations [24]. The existing reviews present a fragmented view of the many facets inherent in understanding the current research landscape of sustainability. Our bibliometric analysis attempts to provide a bird's eye view across the 22 fields in ABS Academic Journal Guide and to provide a more comprehensive assessment of the literature on sustainability from business and management perspective.
Most of the previous reviews used limited numbers of publications and expert-based judgments to evaluate the literature, extract research themes, and suggest directions for future research. For example, Chen et. al [16] used human experts to read the full text of 174 papers on sustainability and supply chain collaboration, the authors read 1126 long-listed articles, and applied their judgments to select 122 articles for review. These reviews are mostly qualitative and are not suitable for understanding the research landscape of a wide, interdisciplinary domain such as sustainability. They also cannot provide dynamic views to illustrate how a research domain has evolved over a long period of time.

Large-Scale Bibliometric Review to Identify Specialties
The ability to take stock of a research field or specialty has increased tremendously in the past decade due to the increasing degree of digitalization of research article databases (e.g., Web of Science, Scopus) and the developments in software tools such as CiteSpace [25] and VOSviewer [26]. Such reviews have shed important light on evaluating the productivity of business schools [27], ranking academic journals [28], stocktaking research domains [29], and discovering emerging research trends and technologies [30].
A number of recent bibliometric reviews have examined the sustainability domain. For example, Cullen [3] reviewed published sustainability research that relates to business and management education over the 20 years (1994-2013). Their review showed a nascent rise in the interest in this domain, but their analysis was primarily descriptive. Chabowski, et al. [31] reviewed the sustainability research in marketing by examining 76,342 citations made in 1320 sustainability-focused articles from 36 journals over 51 years (1958-2008). They employed multidimensional scaling and extracted five broad topics as future research directions. Tang et al. [32] reviewed the ten years of publications in the journal Sustainability since 2009, when it released the inaugural issue. This present review instead covers a time period of 30 years, a larger selection of academic journals (433 journals), and higher-quality publications on sustainability. The review by Zhu and Hua [2] bears the closest resemblance to this present one. They compiled a dataset of 59,926 records and analyzed them using CiteSpace. Yet, their data sample was different from this present review in two ways. First, they did not impose any restriction on the journals and consequently it was a medley of 32 languages, 49 countries, and 149 research areas. Second, their search keyword was "sustainab* development" only, whereas this review, motivated by the multifaceted nature of sustainability, purposely incorporated the search terms related to the 17 SDGs adopted by the U.N. Therefore, we believe this review defines a more suitable sample of existing sustainability research in business and social sciences fields.

The Gap between Academic Research and Practice
Most related literature reviews have explored the humongous research to catch the theoretical trends, but ignore deeply analyzing the gap between academic research and practice [33,34]. As indicated by the empirical evaluation on the usefulness of academic research to practitioners, in particular, business schools' research has limited impact on business practitioners and makes limited contribution for practice, although a number of scholars strive to reduce the research and practice gap [33,34]. Therefore, in this review, centering on the practical 17 SDGs for the sustainability of the U. N., we compared the achieved and untouched topics to find the gap between academic research and practice in Section 5.

Sample
We reviewed sustainability research from 1990 to 2019 from the ISI Web of Science Core Collection database, as adopted by prior research [2,5,35]. Another widely use dataset, Scopus, covers a wider journal range but it is currently limited to recent articles (mainly after 1995) compared with Web of Science [36]. Therefore, we chose the ISI Web of Science Core as the data source. The research articles were collected from academic journals ranked at 4*, 4, and 3 in the Academic Journal Guide 2018, released by the Chartered Association of Business Schools in the U.K. The journals with four stars publish the most original and best executed research papers, and the journals with three stars publish original and well-executed research papers. Although the measurements are different in different countries, these leading journals are generally agreed by scholars in these research fields. "It is argued that there is no perfect method of assessing journal quality, but that the ABS Academic Journal Quality Guide overcomes some of the failings of established methods" [37]. The ABS Academic Journal Guide considers a range of indicators and provides wide journal coverage; has high levels of internal and external reliability; is sensitive to small variations in the ratings of journals; and is generally recognized as a reasonable ways of ranking journals within its user community. It provides a list of journals that cover a wide range of 22 disciplines in business school research and education. As such, it allowed us to investigate the research output in an interdisciplinary and refined way. This journal list is widely adopted by business schools and other stakeholders in evaluating the scientific production and dissemination in business and management disciplines in UK [38]. Collectively, 433 journals in all 22 disciplines in business and management were included based on the five measurements including article citation, institution list, peer surveys, citation studies, and derived lists (Appendix A).
To examine the intellectual landscape, the research domain must be defined clearly. This is rather difficult because the field of sustainability is highly interdisciplinary and involves a wide variety of issues. Our dataset was constructed in three steps. First, we searched for the term "sustainable" or "sustainability" in the titles, keywords, and abstracts of published articles. Second, we added a number of descriptive words that are associated with sustainable development from the United Nations' Sustainable Development Goals including "responsible", "clean", "decent", "affordable", "justifiable", "green", "renewable", "ecological", "environmental", "balanced", and "organic." Third, in business and management, sustainability is concerned with the impacts of business operations on society, impacts of business operations on the environment, equality, diversity, consumer protection, health, environmentally sustainable transportation systems, governmental regimes, environmental risks, information security, information resilience, circular economy, government and sustainability regulation, smart cities and homes, social networks, online sharing, economic growth, employment, income, democracy, social justice, social inclusion, crime, and emerging technologies. Thus, these terms were used in search of the articles as well. We restricted the type of research to be articles and reviews as labeled in Web of Science, since they are predominately original research.
The resultant dataset contained 37,322 published articles as primary citing documents between January 1990 and February 2019. Figure 2 shows the fast increase in the number of research papers published on sustainability relative to the total amount over this period. In 1990 there were only 12 such articles fitting our selection criteria, about 0.098% of the total amount. In 2000, a decade later, it increased to 687 and further to 1717 in 2010, to be 4.39% of all the publications. By 2018, the quantity Sustainability 2019, 11, 5596 6 of 37 of publications on sustainability grew to 3222, to be 10.6% of the total studies, suggesting that the sustainability research has played a more and more important role in business and management research areas. From these publications we generated 1,199,398 secondary documents (i.e., the cited references), which became the basis of document co-citation analysis and clustering. To observe the dynamic evolutions during the past 30 years and detect the evolution over time, we divided the 30 years into three decade-long periods (Table 1) and inspected them closely, as described below in Section 4. of publications on sustainability grew to 3222, to be 10.6% of the total studies, suggesting that the sustainability research has played a more and more important role in business and management research areas. From these publications we generated 1,199,398 secondary documents (i.e., the cited references), which became the basis of document co-citation analysis and clustering. To observe the dynamic evolutions during the past 30 years and detect the evolution over time, we divided the 30 years into three decade-long periods (Table 1) and inspected them closely, as described below in Section 4.  Notes: "Results" column means the total number of searched "Articles" and "Reviews" types of literature. "References" column here represents the total unrepeated cited references. "Authors" and their affiliated "Institutions" column count each author and institution only once in the entire pool.

Analytical Method
This review used two bibliometric methods-document co-citation analysis and dual-map overlay-to study the current research on sustainability published over the past three decades. First, the collections of research papers generated in the steps above were analyzed using document cocitation analysis (DCA). Document co-citation studies a network of co-cited references and describes how two research articles may be linked. They are defined as the "frequency with which two documents are cited together" [39]. The fundamental assumption is that co-citation clusters reveal underlying intellectual structures. For example, Small [39] and Chen [25] are both cited in this paper, thus these two cited references have a co-citation link. With more documents citing these two papers together, the related nature of Small (1973) and Chen (2006) grows stronger. Here, the cited articles are the knowledge base and the citing articles are the research front. The co-citation frequency between two references is calculated based on the number of times that they are cited together in the primary, citing documents.
Such a co-citation analysis serves several functions. First, it maps the relationships between key ideas in a field (Small 1973). The greater the frequency that two papers are co-cited, the stronger relationship there is between the ideas presented in those two works. Second, co-citation analysis identifies landmark articles that are intellectually significant works for a knowledge domain.  Notes: "Results" column means the total number of searched "Articles" and "Reviews" types of literature. "References" column here represents the total unrepeated cited references. "Authors" and their affiliated "Institutions" column count each author and institution only once in the entire pool.

Analytical Method
This review used two bibliometric methods-document co-citation analysis and dual-map overlay-to study the current research on sustainability published over the past three decades. First, the collections of research papers generated in the steps above were analyzed using document co-citation analysis (DCA). Document co-citation studies a network of co-cited references and describes how two research articles may be linked. They are defined as the "frequency with which two documents are cited together" [39]. The fundamental assumption is that co-citation clusters reveal underlying intellectual structures. For example, Small [39] and Chen [25] are both cited in this paper, thus these two cited references have a co-citation link. With more documents citing these two papers together, the related nature of Small (1973) and Chen (2006) grows stronger. Here, the cited articles are the knowledge base and the citing articles are the research front. The co-citation frequency between two references is calculated based on the number of times that they are cited together in the primary, citing documents.
Such a co-citation analysis serves several functions. First, it maps the relationships between key ideas in a field (Small 1973). The greater the frequency that two papers are co-cited, the stronger relationship there is between the ideas presented in those two works. Second, co-citation analysis identifies landmark articles that are intellectually significant works for a knowledge domain. Landmark works are important to examine in bibliometric analysis because they not only reveal the significant contributions made to the research domain at a given point of time, but also indicate intellectual turning points dynamically [40].
This review applies the method of bibliometric mapping, an information visualization technique that can quantitatively display the landscape and dynamic aspects of a knowledge domain [25,41,42]. The mapping tool used here was CiteSpace (v.5.3.R10) (http://cluster.cis.drexel.edu/~{}cchen/citespace/. Accessed 10 February 2019), a Java-based application developed by Chaomei Chen [25,40]. CiteSpace visualizes and analyzes trends and patterns in scientific literature, taking advantage of the citation and co-citation records to present the intellectual structure, including the cluster maps and dual-map overlays, as well as bibliographic coupling. Citespace has become one of the most widely used bibliometric mapping tools [43]. It has been used to map various research areas such as social commerce [44], regenerative medicine [45], hospitality [46], and climate change and tourism [47]. Figure 3 shows the document co-citation network structure over the 30-year period from January 1990 to February 2019. The modularity Q and the mean silhouette score S are two important metrics that tell us about the overall structural properties of the network. The top 100 most cited studies in every slice were applied to generate the references network in that year. We set the time slice to be two years based on trial experiments results, revealing that a two-year time slice is adequate to represent three or more abundant network clusters, among which there are concise relations, representing a sensible performance [29]. The synthesized network contains 37,322 records with 1,199,398 cited references. The network contains 106 co-citation clusters. The 25 largest connected components include 1280 nodes and 1602 links, which account for 82% of the entire network. In this cluster map, the modularity Q of 0.863 is quite high, suggesting that the specialties in sustainability are distinctly defined in terms of co-citation clusters. The mean silhouette score is 0.339, which is acceptable because of the existing numerous small clusters. The silhouette scores of the major clusters we will emphasize in this review are adequately high, as shown in Table A2 of Appendix A. In generating the network, we used the minimum spanning tree algorithm (MST) to prune it to elucidate the network more concisely. Landmark works are important to examine in bibliometric analysis because they not only reveal the significant contributions made to the research domain at a given point of time, but also indicate intellectual turning points dynamically [40]. This review applies the method of bibliometric mapping, an information visualization technique that can quantitatively display the landscape and dynamic aspects of a knowledge domain [25,41,42]. The mapping tool used here was CiteSpace (v.5.3.R10) (http://cluster.cis.drexel.edu/~cchen/citespace/. Accessed 10 February 2019), a Java-based application developed by Chaomei Chen [25,40]. CiteSpace visualizes and analyzes trends and patterns in scientific literature, taking advantage of the citation and co-citation records to present the intellectual structure, including the cluster maps and dual-map overlays, as well as bibliographic coupling. Citespace has become one of the most widely used bibliometric mapping tools [43]. It has been used to map various research areas such as social commerce [44], regenerative medicine [45], hospitality [46], and climate change and tourism [47]. Figure 3 shows the document co-citation network structure over the 30-year period from January 1990 to February 2019. The modularity Q and the mean silhouette score S are two important metrics that tell us about the overall structural properties of the network. The top 100 most cited studies in every slice were applied to generate the references network in that year. We set the time slice to be two years based on trial experiments results, revealing that a two-year time slice is adequate to represent three or more abundant network clusters, among which there are concise relations, representing a sensible performance [29]. The synthesized network contains 37,322 records with 1,199,398 cited references. The network contains 106 co-citation clusters. The 25 largest connected components include 1280 nodes and 1602 links, which account for 82% of the entire network. In this cluster map, the modularity Q of 0.863 is quite high, suggesting that the specialties in sustainability are distinctly defined in terms of co-citation clusters. The mean silhouette score is 0.339, which is acceptable because of the existing numerous small clusters. The silhouette scores of the major clusters we will emphasize in this review are adequately high, as shown in Table A2 of Appendix A. In generating the network, we used the minimum spanning tree algorithm (MST) to prune it to elucidate the network more concisely.

A Landscape of Thirty Years of Sustainability Research Topics
(a) The nodes in the view represent the analyzed subjects (that is, cited references), and the more citations appear, the larger the nodes of frequency. The visualized cluster view represents many useful indicators of the co-citation network. The map can show the formation of knowledge clusters, the association between clusters, and the evolution over time. For example, the connection between nodes represents a co-occurrence (or co-citation) relationship. The color and thickness in the inner circle of the node indicate the frequency (or citations) that occurs in different time periods. The clusters are labeled by noun phrases extracted from the titles, keyword lists, or abstracts of articles that cited the particular cluster. We report the top three labels extracted from keywords using both Latent Semantic Indexing (LSI) [48] and Log Likelihood Ratio (LLR) [49]. Labels extracted by LSI tend to capture implicit semantic relationships across records, whereas those chosen by LLR tend to reflect a unique aspect of a cluster. The areas of different colors indicate the time when co-citation links in those areas appeared for the first time. The change in color from cold to warm tones indicates time changes from early to the more recent. Other nodes with red tree rings are references with citation burst, that is, cited references were suddenly added in the short term [50].
For example, as shown in Figure 3a, the purple-colored areas at the upper right quadrant, including the six labels (#0-liquidity constraints, #12-measurement of diversification, #13-crowdout, #14-gender, #15-difference stationary, #24-agricultural household modles), represent earlier research hotspots. For instance, the #0-liquidity constraints topic, popular in 1990s, means the restrictions on the use of assets to finance needs for consumption or investment, which relates with consumption and income, playing an important role for macroeconomic outcomes. In the middle, the magenta-colored areas represent more recent research hotspots, including #5-financial development, #7-income inequity, #8-unit roots, #9-social capital, #10-happiness, and #17embeddedness. The latest research topic hotspots are covered in the yellow areas, including #1material footprint, #2-trust, #3-panel data, #4-corporate social responsibility, #6-propensity score matching, #16-environmental assessment, #21-consumption income and wealth inequality. A greater number of sudden points can be found in the magenta area and the yellow area, suggesting The nodes in the view represent the analyzed subjects (that is, cited references), and the more citations appear, the larger the nodes of frequency. The visualized cluster view represents many useful indicators of the co-citation network. The map can show the formation of knowledge clusters, the association between clusters, and the evolution over time. For example, the connection between nodes represents a co-occurrence (or co-citation) relationship. The color and thickness in the inner circle of the node indicate the frequency (or citations) that occurs in different time periods. The clusters are labeled by noun phrases extracted from the titles, keyword lists, or abstracts of articles that cited the particular cluster. We report the top three labels extracted from keywords using both Latent Semantic Indexing (LSI) [48] and Log Likelihood Ratio (LLR) [49]. Labels extracted by LSI tend to capture implicit semantic relationships across records, whereas those chosen by LLR tend to reflect a unique aspect of a cluster. The areas of different colors indicate the time when co-citation links in those areas appeared for the first time. The change in color from cold to warm tones indicates time changes from early to the more recent. Other nodes with red tree rings are references with citation burst, that is, cited references were suddenly added in the short term [50].
For example, as shown in Figure 3a, the purple-colored areas at the upper right quadrant, including the six labels (#0-liquidity constraints, #12-measurement of diversification, #13-crowd-out, #14-gender, #15-difference stationary, #24-agricultural household modles), represent earlier research hotspots. For instance, the #0-liquidity constraints topic, popular in 1990s, means the restrictions on the use of assets to finance needs for consumption or investment, which relates with consumption and income, playing an important role for macroeconomic outcomes. In the middle, the magenta-colored areas represent more recent research hotspots, including #5-financial development, #7-income inequity, #8-unit roots, #9-social capital, #10-happiness, and #17-embeddedness. The latest research topic hotspots are covered in the yellow areas, including #1-material footprint, #2-trust, #3-panel data, #4-corporate social responsibility, #6-propensity score matching, #16-environmental assessment, #21-consumption income and wealth inequality. A greater number of sudden points can be found in the magenta area and the yellow area, suggesting a reduced homogeneity in clusters and a Sustainability 2019, 11, 5596 9 of 37 proliferation of research topics. Such a developing process can also be clearly viewed by the Figure 3b, just like a process of mycelium cultivation.

The Decade-By-Decade Analysis
To compare the research topic changes in the sustainability domain, research topics in Periods I, II, and III, defined by the 1990s, 2000s, and 2010s, respectively, were used.  Tables 2-4 show the co-citation clusters identified in the network. Each of the largest 6 clusters of these three periods all have more than 100 members. Their silhouette scores indicate a high level of homogeneity within these clusters. For each of the 10-year periods, we focused on the largest topic clusters to illustrate the research fronts in different decades over the 30 years. The largest connected components include 373 nodes, which account for 71.5% of the entire network. In this cluster map, the relatively high modularity Q of 0.858 suggests the clustering is reasonable. The mean silhouette score of 0.448 suggests that the homogeneity of these clusters, on average, is moderate. The largest 11 clusters all have more than 20 members each as shown in Table 2.
Sustainability 2019, 11, x FOR PEER REVIEW 9 of 39 a reduced homogeneity in clusters and a proliferation of research topics. Such a developing process can also be clearly viewed by the Figure 3b, just like a process of mycelium cultivation.

The Decade-By-Decade Analysis
To compare the research topic changes in the sustainability domain, research topics in Periods I, II, and III, defined by the 1990s, 2000s, and 2010s, respectively, were used.  Tables 2-4 show the co-citation clusters identified in the network. Each of the largest 6 clusters of these three periods all have more than 100 members. Their silhouette scores indicate a high level of homogeneity within these clusters. For each of the 10-year periods, we focused on the largest topic clusters to illustrate the research fronts in different decades over the 30 years. The largest connected components include 373 nodes, which account for 71.5% of the entire network. In this cluster map, the relatively high modularity Q of 0.858 suggests the clustering is reasonable. The mean silhouette score of 0.448 suggests that the homogeneity of these clusters, on average, is moderate. The largest 11 clusters all have more than 20 members each as shown in Table 2.   Note: Clusters #2 to #4 were pruned and not reported in the CiteSpace output due to their extremely low centrality. Figure 4 consists of two rather concentrated sections in its upper and lower parts, respectively. Occupying the upper part is the largest cluster in this period, focusing on household consumption and economic security as represented in "consumption", "liquidity constraints", and "permanent income" [51]. The next largest cluster is concerned with economic growth in macroeconomics [52], followed by the third largest cluster concentrating on topics such as job mobility, innovation, and labor markets, such as agriculture and the youth labor market [53]. They are clustered in the bottom of Figure 4. For this period, the research, represented in the publications in the ABS Academic Journal Guide 2018 prestigious journals, have mostly involved the economic factor as the single pillar of sustainability.

Period II (2000-2009)
Research topics in the first decade of the 21st century are presented in a clustered landscape view in Figure 5. Again the 10-year period is sliced at two years and the top 100 most cited studies in each slice were employed to generate the references network in that slice, containing 10,848 records with 323,222 references. Fifty-two clusters were formed, with the largest connected components consisting of 347 nodes out of 430 nodes, accounting for 80.7% of the entire network. The modularity Q of 0.824 is relatively high, suggesting that the network is reasonably divided into loosely coupled clusters.
The cluster map has a moderate mean silhouette score of 0.366, indicating the existence of numerous small clusters. We again focus on the largest clusters to understand the research landscape in this period, summarized in Table 3.

Period II (2000-2009)
Research topics in the first decade of the 21st century are presented in a clustered landscape view in Figure 5. Again the 10-year period is sliced at two years and the top 100 most cited studies in each slice were employed to generate the references network in that slice, containing 10,848 records with 323,222 references. Fifty-two clusters were formed, with the largest connected components consisting of 347 nodes out of 430 nodes, accounting for 80.7% of the entire network. The modularity Q of 0.824 is relatively high, suggesting that the network is reasonably divided into loosely coupled clusters. The cluster map has a moderate mean silhouette score of 0.366, indicating the existence of numerous small clusters. We again focus on the largest clusters to understand the research landscape in this period, summarized in Table 3.      Although the period continues to demonstrate a strong dominance of economics in the publications in the ABS Academic Journal Guide selected journals consistent with previous findings in Azar [54], the largest clusters show remarkable shift from the previous decade. In particular, three of the largest clusters (Clusters #0, #1, and #3) are characterized by research on socioeconomic factors and social capital and how social network factors affect individual health [55]. Research in psychology [56] and management [57] have become prominent in informing the research in this decade, especially in providing methods and conceptualization in understanding the effects of social factors. Cluster #2 represents an interesting shift to subjective well-being and happiness [58] and not just the income and consumption of individuals and households, as in Period I. These research topics form the large concentration of clusters on the right-hand side of Figure 5.
The left-hand side of Figure 5 shows that topics related to economic growth continue to be a large mass of clusters but their prominence has reduced (Clusters #4, #6 and #7). The attention of economists in this decade converges with the clusters on the right-hand side. Instead of focusing on general equilibrium in macroeconomics and stable economic growth, at the beginning of the 21st century, economists were looking into inequality and its various causes and consequences in economic growth such as income inequality-induced health inequality [59] and imbalance in world economic growth caused by ethnic division [60]. Institutional factors that may affect cross-nation differences in economic growth and stability, such as democracy [61], have gained attention. In summary, the research published in these ABS Academic Journal Guide prestigious journals in Period II has been enriched to incorporate a second pillar of sustainability, namely, social sustainability. To analyze the research topics in Period III, we again generated the cluster map by identifying the top 100 articles of each subsequent 2-year long slice. The modularity Q is 0.793 and the mean silhouette is 0.482. The generated network contains 405 nodes out of the total of 421, roughly 96.2%. The research topics have greatly diversified visibly ( Figure 6) and there is no clear division among the clusters. The cluster labels summarized in Table 4 confirm the interdisciplinary nature of this period of business research. The largest cluster (Cluster #0) builds on econometric methods such as panel data analysis [62,63], estimation of nonlinear models, and generalized method of moments estimation [64]. The interesting new development is evident in the center of Figure 6, where it shows the second largest cluster (Cluster #1) "happiness". This cluster includes a number of research papers which employ methods such as propensity score matching to draw causal inference [65] and apply them to assessing the effects of policy interventions and other treatments of interest on happiness and subjective well-being [66]. Cluster #2 demonstrates the change in technology and the research interests shifting to social media [67], innovation [68], and big data [69]. Table 5 juxtaposes the top 10 topics in the sustainability field from each of the three periods to highlight the shifts in focus. The three pillars of sustainability framework suggests that sustainability comprises not just economic development, but also social development and environmental protection. The three decades showed a clear shift from a nearly exclusive focus on economic growth and consumption to all three pillars. For the period I, the leading studies have mostly involved the economic factor as the single pillar of sustainability; Period II was enriched to incorporate a second pillar of sustainability, namely, social sustainability; in Period III, emerging topics such as environmental assessment, environmental management, and corporate social responsibility have played important roles in the third pillar of sustainability, namely environment.  Figures 7-9, respectively, visually present the timeline views of research topics year by year in each of the 10-year periods studied in this review, where each cluster is arranged on a horizontal timeline. The direction of time points to the right in the timeline view, hence the color runs from cold (distant past) to warm (recent past). The clusters labeled with descending size are placed vertically on the right side. Each year the top 100 most cited references in each slice are shown along each of the timelines. From the figures, it is easy to inspect the temporal patterns of how the clusters of co-cited references evolved over time by extracting the year-by-year labels identified by the LSI method. The timeline on the top is the largest cluster (Cluster #0). Each node represents a cited reference. The tree rings represent citations received by the reference over the years (inside out). The purple ring of a node indicates its strong betweenness centrality. The red rings indicate a period of citation burst. Labels show representative terms used by articles that cited a cluster in a particular year. The red core of a node indicates there was a citation burst in earlier years of the period. The co-citation links appended in the year of the corresponding color are displayed by the colored curves [70]. Figures 7-9, respectively, visually present the timeline views of research topics year by year in each of the 10-year periods studied in this review, where each cluster is arranged on a horizontal timeline. The direction of time points to the right in the timeline view, hence the color runs from cold (distant past) to warm (recent past). The clusters labeled with descending size are placed vertically on the right side. Each year the top 100 most cited references in each slice are shown along each of the timelines. From the figures, it is easy to inspect the temporal patterns of how the clusters of cocited references evolved over time by extracting the year-by-year labels identified by the LSI method. The timeline on the top is the largest cluster (Cluster #0). Each node represents a cited reference. The tree rings represent citations received by the reference over the years (inside out). The purple ring of a node indicates its strong betweenness centrality. The red rings indicate a period of citation burst. Labels show representative terms used by articles that cited a cluster in a particular year. The red core of a node indicates there was a citation burst in earlier years of the period. The co-citation links appended in the year of the corresponding color are displayed by the colored curves [70].     From the first 10-year period to the blue-colored areas of the second one, there is significant change in the topics that qualify as a citation burst, especially the social capital and various kinds of inequality that have taken up much attention of scholars. There were many more hotspots in Period II, indicating a time of transition. In the last 10-year period in our data, happiness has become a new research hotspot. Environmental assessment and environmental management (especially in supply chain management) have gained prominent positions. The topics of inequality repeatedly appear in multiple clusters (Clusters #3 and #8). The shift from economic to societal to environmental nearly perfectly affirms the "three pillars" framework of sustainability. From the first 10-year period to the blue-colored areas of the second one, there is significant change in the topics that qualify as a citation burst, especially the social capital and various kinds of inequality that have taken up much attention of scholars. There were many more hotspots in Period II, indicating a time of transition. In the last 10-year period in our data, happiness has become a new research hotspot. Environmental assessment and environmental management (especially in supply chain management) have gained prominent positions. The topics of inequality repeatedly appear in multiple clusters (Clusters #3 and #8). The shift from economic to societal to environmental nearly perfectly affirms the "three pillars" framework of sustainability.

Timeline
The duration, sustainability, and activeness for a cluster to remain interesting are shown in Table 6. For example, happiness research had remained active for 14 years, but this topic has become relatively quiet since 2015. In contrast, corporate governance, with a similar longevity of 14 active years, is still active, although its prime time was until 2017. Inequality and economic growth are evergreen research topics and this bibliometric analysis confirms that they will remain active as the cornerstones of sustainability. Corporate social responsibility (CSR) as a research topic became active in 2009 [71] and has remained active. Figure 10 presents a micro-targeted view of cluster #5, corporate social responsibility, which shows the evolution of this topic and the articles that the citation burst is associated with. From the first 10-year period to the blue-colored areas of the second one, there is significant change in the topics that qualify as a citation burst, especially the social capital and various kinds of inequality that have taken up much attention of scholars. There were many more hotspots in Period II, indicating a time of transition. In the last 10-year period in our data, happiness has become a new research hotspot. Environmental assessment and environmental management (especially in supply chain management) have gained prominent positions. The topics of inequality repeatedly appear in multiple clusters (Clusters #3 and #8). The shift from economic to societal to environmental nearly perfectly affirms the "three pillars" framework of sustainability.
The duration, sustainability, and activeness for a cluster to remain interesting are shown in Table  6. For example, happiness research had remained active for 14 years, but this topic has become relatively quiet since 2015. In contrast, corporate governance, with a similar longevity of 14 active years, is still active, although its prime time was until 2017. Inequality and economic growth are evergreen research topics and this bibliometric analysis confirms that they will remain active as the cornerstones of sustainability. Corporate social responsibility (CSR) as a research topic became active in 2009 [71] and has remained active. Figure 10 presents a micro-targeted view of cluster #5, corporate social responsibility, which shows the evolution of this topic and the articles that the citation burst is associated with.

Business Research in the Journal of Sustainability
The core value of Sustainability journal is openness to diversity based on a variety of disciplinary lenses about sustainability. Here, we also map the 955 related articles in the journal from 2009 to 2019 and find some interesting points for reference, as shown in Figure 11 and Table 7. The topics with the orange color are closely interconnected with and transformed from the topics of the yellow color gradually. The business model innovation has been the most outstanding topic [72], followed by topics of circular economy [73] and group decision making. Sustainability in developing countries has been paid more attention and strategic orientation for the management in sustainability has been proposed. The classic topics like corporate responsibility and family firms are still developing. The OHSAS 18001 Certification Scheme, an occupational health and safety standard (OH&S) designed to enable organizations to control risks and improve performance in the area of OH&S, has been recently increasing in promotion. It is noticeable that these more specified topics extracted from the journal of Sustainability are same as the topics we refined from the past few decades, such as social responsibility, while also different from many topics, such as panel data, happiness, environmental assessment, and so on. These nonoverlapping topics can shed light on the readers and scholars of the potential extended research topics of business and management for sustainability in the journal.  The core value of Sustainability journal is openness to diversity based on a variety of disciplinary lenses about sustainability. Here, we also map the 955 related articles in the journal from 2009 to 2019 and find some interesting points for reference, as shown in Figure 11 and Table 7. The topics with the orange color are closely interconnected with and transformed from the topics of the yellow color gradually. The business model innovation has been the most outstanding topic [72], followed by topics of circular economy [73] and group decision making. Sustainability in developing countries has been paid more attention and strategic orientation for the management in sustainability has been proposed. The classic topics like corporate responsibility and family firms are still developing. The OHSAS 18001 Certification Scheme, an occupational health and safety standard (OH&S) designed to enable organizations to control risks and improve performance in the area of OH&S, has been recently increasing in promotion. It is noticeable that these more specified topics extracted from the journal of Sustainability are same as the topics we refined from the past few decades, such as social responsibility, while also different from many topics, such as panel data, happiness, environmental assessment, and so on. These nonoverlapping topics can shed light on the readers and scholars of the potential extended research topics of business and management for sustainability in the journal.

Research Gaps and Insights
In Table A3 (Appendix A), we compared and listed the differences between goals of the practitioners (e.g., SDGs of the UN) and the research directions taken by academia in order to understand if research has caught up with fulfilling practical needs and postulate the reasons why research work have (or have not) adequately done so. These insights could then be used to synthesize the research gaps and best practices for sustainability researchers and practitioners.
We find that there is one category of goals to which the business and management scholars have paid attention, but it needs to be further explored in correlation with current circumstances; while another category of goals are still to be explored further with consideration of the compatible interesting research points. Category I goals that scholars have been gradually considering include various research topics on goals more related with business and management, such as good health and well-being; decent work and economic growth; industry, innovation, and infrastructure; sustainable cities and communities; reduced inequalities; and responsible consumption and production. While category II demonstrates some goals more about natural science that still need to be further explored from business and management perspectives, such as clean water and sanitation, affordable and clean energy, climate action, life below water, and life on land. One obvious reason for the lack of studies in category II is that the cross-disciplinary barriers between business management and these domains, such as hydraulic engineering, energy production, climatology, marine science, biology, and geography, are higher. However, sustainability intrinsically refers to the fusion of these different specialties; closer multidisciplinary cooperation in academic areas has been a trend that can break these cross-discipline barriers. The tricky point is how to break the boundaries and fill up the gap by exploring the valuable research questions and providing solutions from the perspectives of business and management. To deal with this, following the clues of these sustainable development goals, we first classified the related topics in three different periods that have contributed to these goals, as shown in Table A3 (Appendix A), in order to find the research gap. Then, these research gaps for the goals were further explored through the latest enlightening literature from the influential journals in recent years, which led to new insights in the following subsections. Besides, another perspective to link the multidisciplinary sustainability problem with business and management is the innovative consideration of applying the management theories or approaches to provide sustainable solutions. For example, to achieve the SDGs in Category II that could be seen as out of the reach of firms (and for this reason not much explored by the business and management literature), we can consider this through firms' efforts to establish partnerships to enhance policy dialogue and coordination to achieve sustainability targets. Such an innovative perspective can be applied for each of the SDGs to generate meaningful studies.
As Table A3 (Appendix A) shows, the main topics explored by Citespace in each period have been enlisted. Each topic belongs to a cluster, as shown in Tables 2-4, denoted as cluster numbers in the parentheses. The smaller the cluster number, the more attractions on this topic in the specific period. For instance, in Table A3, the food security (0) topic is mined in Period II, related with the zero hunger goal. This food security has been one of the most concerned topics for scholars in Period II.

Good Health and Well-Being
Ensuring healthy lives and promoting well-being at all ages is essential to sustainable development. Currently, health care has been a fast-growing research area attracting more scholars in the business and management area, especially in operations research and information science. The following practical objectives can enlighten the scholars: mortality ratio, diseases, traffic safety, sexual and reproductive healthcare services, health financing, health workforce management, and health risks management. All these proposed topics can be viewed from different perspectives based on the demands of stakeholders, as shown in Table 8. It is also worth noting that few studies on traffic safety have been published in these influential journals in recent years. With numerous fatal tragedies caused by road and air traffic safety problems, the related traffic safety topics cannot be overemphasized with topics such as traffic control policy, aviation safety, human-machine interaction, self-driving safety, and so on. • Health workforce management considering stress, efficiency and productivity; Health risks management Health information technology [75] • Health risk management with IT; emergency management;

Decent Work and Economic Growth
As the fundamental part of the three pillars of sustainability, economic growth is the most prosperous research area. Sustainable economic growth will require societies to create the conditions that allow people to have quality jobs that stimulate the economy without harming the environment. Up to now, there has been abundant literature about the goals of decent work and economic growth. In measuring the gap between the status quo and the goal, as shown in Table 9, several topics can be studied, including decent job creation; entrepreneurship; the formalization and growth of micro-, small-, and medium-sized enterprises; safe and secure working environments; and sustainable tourism that creates jobs and promotes local culture and products. Among these topics, the decent job creation and entrepreneurship topics are both related with new economic forms such as "gig economy" or "sharing economy". Besides, the innovative ways for employment, such as self-employment, and financing, such as crowdfunding, also spring up with the development of IT technology. Further, more attention can be put on the topic "all women and men receive equal pay for work of equal value", dealing with the existing gender income gap, for whichmore approaches are still waiting to be proposed. These provide interesting insights for future research in the related topics. Finally, the sustainable tourism that creates jobs and promotes local culture and products is also an interesting topic that is worthy of further exploration.

Industry, Innovation, and Infrastructure
This goal is to build resilient infrastructure, promote sustainable industrialization, and foster innovation. The past few decades have witnessed a great progress of industry, innovation, and infrastructure-which are crucial to achieving sustainable development and empowering communities in many countries. For the next phase, as shown in Table 10, we pursue the sustainable and resilient infrastructure, industrialization, clean technologies, improvement of the ICT in undeveloped countries, and promote innovation by increasing the R&D workers. Technological progress is the foundation of efforts to achieve environmental objectives, such as increased resource and energy-efficiency. Without technology and innovation, industrialization will not happen, and without industrialization, development will not happen. There needs to be more investment in high-tech products, such as 5G, AI, deep learning, self-driving, and blockchain, that dominate the manufacturing productions in order to increase efficiency and a focus on mobile cellular services that increase connections between people. Innovation and increasing R&D workers Academic scientist mobility for industrial innovation [111] •

R&D workers
Information and communication technology Internet use and mobile [112] Digital service innovation and bridging the service divide [113] Sustainability of polycentric information commons [114] Ecosystem of software platform [115] ICT in developing country (India) [116] • ICT in undeveloped area; • Sustainable IT; • Mobile and Internet; • Sustainable platform ecosystem;

Reduced Inequalities (Goal 10) and Gender Inequalities (Goal 5)
The unbalanced development widely exists within and across countries, therefore, the reducing economic inequality goal is set to reduce inequality that persists and the large disparities that remain regarding access to health and education services and other assets. Another goal for inequalities is gender inequality. This goal is set to achieve gender equality and empower all women and girls.
As represented in Table 11, reducing gender inequality has been studied with consideration of firm role [117], training [118], start funding [96], CSR [119], and hiring [120]. More inequalities exist in age, disability, race, ethnicity, origin, religion, economic, or other status and the research gap still exists in enhancing the use of enabling technology, in particular information and communications technology, to promote the empowerment of women and end gender discrimination, such as the online labor market [121] and crowdfunding [93]. Table 11. Research gap, enlightening literature and insights for reduced inequalities (Goal 10) and gender inequalities (Goal 5).

Gap Enlightening Literature Insights
All women and men equal pay for work of equal value Firm role [117]; Training [118]; Gender gap in start funding [96] Gender diversity and CSR [119] Hiring preferences in online labor markets [120] • Gender gap under different circumstances; • Ways to reduce the gender gap such as internship and training; Inequalities in age, disability, race, ethnicity, origin, religion or economic or other status Racial bias and interaction [121] Team performance and racial bias [122] • Inequalities in different aspects besides race, such as age, disability, ethnicity, origin, religion or economic or other status.
Income inequality Technology innovation and executive pay inequality [123] Race bias in Crowdfunding [93] Double standards [124] • Income inequality under IT context.

Peace, Justice, and Strong Institutions
Since the beginning of the new millennium, there has been a development of multipolarization of the world, economic globalization, informatization, and cultural diversification, and despite the fact that the trend of peaceful development is strong, ethnic conflicts and cross-border disputes are still frequent; extremism and terrorism is unpredictable; information crimes are rampant; the damage to peace is more uncertain in breadth and depth than ever before; and the sources of threats are more complex and diverse than ever. As shown in Table 12, there are several potential points that can be further explored, and the influence of political factors on stakeholders in business management, considering the potential political risk and uncertainty context, and the new forms of crimes through high-tech or social networks are impeding questions. The dark side of the new media through technology, such as cyberbullying, addictive use, trolling, online witch hunts, fake news, and privacy abuse, are interesting points that merit exploration.

Gap Enlightening Literature Insights
The influence of political factors on stakeholders in business management Political ideology shape consumer's preferences for differentiation [125,126] Managers' political ideology [127] State ownership and firm innovation [101] • Promote and enforce non-discriminatory policies;

Natural-Environmental Goals
As we mentioned before, because of the cross-discipline barriers, the category II for some goals, such as clean water and sanitation, affordable and clean energy, climate action, life below water, and life on land, haven't been sufficiently explored from a business perspective. Based on the extant literature, we probe into this from three points, i.e., environmental management policy, climate change influence, and business in new energy industries, as Table 13 shows. First, government plays an important role in the environmental management process. In agreement with the last partnership of goals, the cross-sector partnerships and endeavors for implementing environmental protection policies are worth being studied. Then, with environmental degradation such as climate change, the emergency management plans for dealing with natural disasters is also an important spot. Finally, the renewable energy operation management issues such as storage, use, and pricing, are all potential research topics for logistics and supply chain management. • Optimization the (renewable) energy location, storage, and investment using modeling methods

Conclusions
Due to the fact that the notion of sustainability is multifaceted, the United Nations charted a total of 17 Sustainable Development Goals for 2030. Academic research on sustainability, as a result, naturally involves multiple disciplines in business, management, and social sciences. Focusing on the large-scale literature from high-quality journals ranked by ABS Academic Journal Guide 2018, our study retrieved sustainability-related articles from Web of Science for analysis and visualization.
We found that the body of research in business and management has evolved greatly when it comes to sustainability. A closer look at the decade-by-decade evolution indicates that research topics have moved from a dominant focus on economic growth to incorporating social factors such as social capital and social networks, and further to paying more attention to a much wider range of issues, especially those related to environment and ecology. Interestingly, happiness has become a new research hotspot in the last 10-year period in our data. Besides, some evergreen topics such as corporate governance, inequality and economic growth, and corporate social responsibility, are likely to remain active in the future. Moreover, environmental assessment and environmental management (especially in supply chain management) have gained noticeable positions. In all, this pattern of evolution corresponds to the three pillars of sustainability: economic, societal, and environmental sustainability, gradually emphasizing more of the importance of the environmental aspect.
In addition, we discussed the emerging topics in the sustainability research field and mining of the research gaps and insights following the Sustainable Development Goals. We compared and listed the differences between goals of the practitioners (e.g., SDGs of the UN) and the research directions taken by academia in order to understand if academic work has caught up with fulfilling practical needs, and postulate the reasons on why research works have (or have not) been done adequately. Seven goals related to business and management were further mined from recent prestigious journal publications to provide the innovative and significant perspectives (as shown in Tables 8-13) for future research and industrial practices. (1) For the goal of good health and well-being, popular research topics include hospital operation management to reduce mortality ratio, online and offline health care for treating diseases, traffic safety, digital technology to improve sexual and reproductive health-care, health workforce management, health financing, and health risks management. (2) There are abundant innovative research points for pursuing decent work and economic growth. Examples include AI technology effects on job creation, online labor market, gig economy, sustainable tourism, emerging markets, and so on. (3) To achieve the growth of the industry, innovation, and infrastructure goal, more studies considering resilient IT infrastructure, new IT technology, and sustainable platform ecosystems are highly needed. (4) With limited amounts of studies on inequalities, it is still urgent to study how to reduce inequalities in age, disability, race, ethnicity, origin, religion, economic, or other statuses under different business circumstances, especially within the information technology innovations context. (5) The goal of peace, justice, and strong institutions shows significant potential for exploration. Many interesting points can be interrelated and explored, such as promoting and enforcing nondiscriminatory policies, customers' preferences affected by politics, managers' decision making influenced by politics, political risk and uncertainty considerations, crime and high-tech, and privacy concerns. (6) The potential research gaps in sustainable consumption and production include: sustainable consumers' behaviors and characteristics in new technological environments such as online platform; supply chain sustainability; and sustainable behavior stimulating policies. (7) The significance and importance of environment-related studies can never be overemphasized from the perspective of business and management. For example, the effective environmental policies such as cross-organizational cooperation, transnational climate policy, nonprofit organization management, and emergency management for disasters can be further explored; the relations between corporate and environmental sustainability, the operations research for renewable energy storage, utilization, and investment are also significant research directions for both society and enterprises. To sum up, for scholars, these topics are like treasures waiting to be discovered, especially for business and management researchers who are interested in multidisciplinary issues related to sustainability. For practitioners, these topics reveal the existing or potential threats and problems for business practice. The employees, employers, entrepreneurs, governmental officials, and the stakeholders, everyone can attempt to explore the ways to take responsibility to take action for sustainability. Served for soliciting the valued opinions of future studies, these topics also illuminate what the practitioners can do to facilitate sustainable development to attain the SDGs, for example, encouraging the policies for reducing bias and inequalities, adapting new information technologies appropriately, taking green consumption, development to control emissions and pollution levels, and so on. Therefore, both the academics and practitioners can use the insights from our analyses to synthesize the current research gaps and best practices.