Identifying Research Trends, Active Research Areas and Pivotal Publications with Co-citation Analysis in CiteSpace: A Case Study with Active Matter

Yuening Zhang
STEM Librarian
University Libraries
Kent State University
Kent, OH
yzhan121@kent.edu

Abstract

Keywords: Research trend analysis, Active research areas. Big data analysis, Co-citation analysis, Research support, Research impact, Bibliometrics, CiteSpace, Active matter, Data visualization

Recommended Citation:

Zhang, Y. (2025). Identifying research trends, active research areas and pivotal publications with co-citation analysis in CiteSpace: A case study with active matter. Issues in Science and Technology Librarianship, 109. https://doi.org/10.29173/istl2779

Introduction

As universities compete for talents and limited research funding, academic libraries are playing a vital role in this process through research support. A study analyzed the websites of 79 academic libraries of the top 100 universities based on the QS World University Rankings (https://www.topuniversities.com/) (Si et al., 2019). They divided research support services into seven groups: “research data management, Open Access, scholarly publishing, research impact, research guides, research consultations, and research tools recommendation.” In the “research impact” categories, services included research metrics (journal impact factors, the h-index, and altmetrics), author identifiers (ORCID and ResearcherID) and other tools to measure research output. A survey of all eight New Zealand university libraries revealed four main research support services areas: research impact, research data management, scholarly communication and indigenous people research (Howie & Kara, 2022). Their study indicated that, between 2018 and 2012, more academic libraries in New Zealand collaborated with other university units to deliver research impact services in areas such as citation reports, calculation of research impact and disciplinary research trends report. A web content analysis of twelve Chinese universities demonstrated that research impact services in major Chinese academic libraries have been elevating from traditional services such as citation analysis to higher level services such as subject/field trend analyses and talent evaluations (Ye, 2019). It is thus more important for academic librarians to be able to do research trend analysis and identify research frontiers and influential publications and authors with bibliometrics tools.

There are several major bibliometrics software for big data analysis such as CiteSpace, VOSviewer (Van Eck & Waltman, 2010), and Sci2 (Lewis & Alpi, 2017). CiteSpace is a powerful java-based bibliometrics tool for complex network analysis with co-citation network analysis as its core function (Chen, 2006).

Citations and co-citations have been well studied in information science to identify influential papers, authors, and journals. For example, Garfield’s algorithmic historiography and HistCite software depicted citation networks (Garfield et al., 2003). On the other hand, Co-citation was defined as “the frequency with which two items of early literature were cited together by the later literature” (Small, 1973, p.265). Small and Griffith (1974) first introduced co-citation networks to reveal science specialties. Small and Sweeney (1985) further developed this technique by introducing “fractional citation counting” and “variable level clustering.” Braam et al. (1991) used co-citation networks to study how scientific activities changed with time. Persson (1994) analyzed articles published from 1986-1990 in Journal of the American Society for Information Science with co-citation networks of authors and demonstrated that this technique could represent the intellectual structure of a field. As demonstrated in Figure 1, article A cited articles 1, 2 and 3, and thus articles 1, 2 and 3 were co-cited by article A. Similarly, articles 3, 4 and 5 were co-cited by article B. In CiteSpace, co-citation articles were grouped into clusters of homogeneity based on types of analysis (co-citation, keywords co-occurrence, or co-authorship), which was considered as the knowledge base of a research domain. In Figure 1, co-cited articles 1, 2 and 3 were from the same cluster, whereas co-cited articles 3, 4 and 5 were from three different clusters. In CiteSpace, cluster names were extracted from the title, keywords or abstracts of citing articles (A and B) in order to identify research trends/fronts. Nodes in the knowledge base represent individual publications and links between them indicate co-citation occurrences.

There are three types of nodes to identify potentially important publications in the co-citation networks (Chen, 2004), as demonstrated in Figure 2. A landmark node is one with a large size, indicating high citation counts. A hub node demonstrates a high degree, which means high co-citation occurrences in the network like a transportation hub. A pivot node is a unique feature in CiteSpace, which functions as a bridge and the shortest path connecting different clusters (Yang et al., 2017). It is measured with betweenness centrality and serves as an indicator of pivotal or intellectually turning point publications (Chen, 2005). Another unique feature of CiteSpace is the detection of publications with citation bursts based on Kleinberg’s algorithm (Kleinberg, 2002), which means a surge in citations and thus indicates active research areas.

CiteSpace has been used in the areas of health sciences (Sabe et al., 2023), education (Wei et al., 2023) and environmental science (Zhang et al., 2023) to identify the evolution of research trends in each field. There also have been a handful of reports of its use in areas of science and engineering, such as in Kansei engineering (Lin et al., 2021), additive manufacturing (Wang et al., 2022), and computer applications in engineering education (Rawat & Sood, 2021). However, it appears that the above analysis of research idea evaluation and trends analysis were all written by researchers in those areas. In this case study, active matter was used as an example to demonstrate how to identify research trends, hot spots and pivotal publications using CiteSpace without subject expertise . Active matter refers to collective pattern formation and self-organization in systems composed of self-propelled units, such as flocks of birds or non-living objects, which can harvest energy and information from their environment (Sugi et al., 2021). It started with Vicsek’s theoretical model (Vicsek et al., 1995) and has been a rapidly growing research area in recent years of physics, biology and material science with applications in molecular motors, robotics, nano-medicine and biomimicking materials (Needleman & Dogic, 2017).

Literature Review

Academic libraries research support services align with research cycles and reflect different phases of scientific communication. It demonstrates an institution’s research output quantity and quality that can influence government funding decisions and university rankings (Wilson, 2017). A recent survey and interviews with faculty members at a U.S. public university discovered that requests for increased research support was the third most common theme out of nine total (Miller et al., 2023).

There are various types of research support services. Fernández-Marcial et al. (2016) analyzed library services from top ten universities on the Academic Ranking of World Universities of the University of Shanghai, and the Times Higher Education World University Rankings. Their study revealed that Open Access and research data were two most developed research services.

The importance of research assessment led academic libraries in the UK and Australia to provide support in this area. The Research Impact Measurement Service (RIMS) in an Australian university was launched in 2007 and over the last decade has shifted from producing research impact reports to an advisory role (Drummond, 2014). Their reports included measurements such as the h-index, citation counts, journal impact, and research trends, which provided valuable information in grant applications, faculty promotion, and institutional benchmarking (Drummond & Wartho, 2009). In their advisor role, they delivered a series of workshops on managing researcher profiles, strategic publishing, and tracking scholarly impact that served both early career and established researchers (Drummond, 2014). A survey of 140 academic libraries from Australia, New Zealand, Ireland and the UK suggested that there had been increasing involvement but there was still significant growing space in library support for trend analysis, faculty review, and talent evaluation (Corrall et al., 2013). An interview of a Chinese medical university’s faculty and librarians indicated that research trend analysis, identification of influential authors and institutions, and collaboration network analysis were desired library research support (Chen & Zhou, 2021). However, lack of subject and/or disciplinary knowledge was identified as a factor that affected research impact service (Corrall et al., 2013).

Bibliometrics has long been used by information professionals and scientists to analyze research impact, which saved time and made the process quantifiable. Moral-Muñoz et al. (2020) divided bibliometric software into three categories: general bibliometric performance analysis tools (such as Publish or Perish), science mapping tools (such as SCI2, CiteSpace and VOSviewer), and Python and R libraries (such as Bibliometrix). On the other hand, a study surveyed 30 world top university library websites for data visualization services, and revealed that Tableau (63.33%), R (56.67%), Excel (40%), Gephi (40%) were top popular tools (Zakaria, 2021). Another study selected ten science mapping tools and concluded that CiteSpace, HistCite and VOSviewer were the most frequently mentioned tools in Web of Science (WOS) publications (Pan et al., 2018). They also reported that the use of these three tools had been increasing, with VOSviewer being mentioned more often than the other two tools.

Science mapping tools originated in the fields of library and information science, and computer science, and were applied in many other fields such as business, physics, chemistry, education, biology, and nursing, to name just a few (Pan et al., 2018). In library and information science, document co-citation with CiteSpace has been applied to identify emerging trends from 2009 to 2016 (Hou et al., 2018). Very recently Biblioshiny R and VOSviewer were used to identify influential authors, major contributing countries, and co-authorship networks of publications on the application of artificial intelligence in libraries (Borgohain et al., 2024). Science mapping tools have also been applied by librarians in research support services. Wu and Tsai (2022) discussed the use of HistCite in bibliometric analysis, including pre-processing of data and export of data to Gephi or Pajak for data visualization. However, HistCite has ceased updates. Lewis and Alpi (2017) provided a guide on how to use SCI2 (Science of Science) to analyze co-authorship, author-paper, citation and journal bibliographic coupling networks as a tool for research support and collection development. However, it seems SCI2 also ceased updates since 2018. There also have been case studies on the use of VOSviewer for librarians. MacDonald and Dressler (2018) demonstrated how to identify research fronts in the “Internet of things” field with VOSviewer in support of library services to tech-transfer. McAllister et al. (2022) introduced how to track research interest shift and reveal collaboration networks by using VOSviewer to analyze publications from science and engineering departments of an institution. However, it appears that there have no case studies on the use of CiteSpace, one of the top three most widely used science mapping/bibliometric tools, to demonstrate how it can be used by librarians to provide data-informed research impact services without subject knowledge.

In this current case study, the author demonstrated how to use a co-citation bibliometric tool to identify research fronts and influential publications with a step-by-step methodology and how to interpret the major structural indicators without subject knowledge. At the end of the analysis, top pivotal publications identified were evaluated by two faculty in this area. Through the introduction of the co-citation analysis techniques with CiteSpace, the following questions will be answered:

• How can librarians learn research topics in a field beyond identification of a few keywords?
• How can we see the shift of research interests in a field over time?
• What additional information can local citation counts provide compared with that from global citation counts?
• How can the research fronts and active research areas be identified?
• How can influential publications in a field be recognized beyond citation counts?

This study illustrates how to provide a research trend and impact analysis off the ground without subject expertise. To the author’s knowledge, case studies on other major science mapping/bibliometric tools for librarians typically focused on shift of keywords over time and author collaboration networks, and did not address the above listed research questions.

Methodology

Active matter was selected as the example of an emerging research field based on discussions with a physics faculty member and reviewed literature (Hallatschek et al., 2023; Needleman & Dogic, 2017). The WOS Core collection was chosen as the bibliographic database because data downloaded from it is the default data source in CiteSpace and can be processed directly by the software after duplicate removal. Bibliographic records were collected from the Web of Science Core Collection on February 20th, 2023, with a topic search of the following five phrases with all date range:
"active matter" OR "active systems" OR "self-propelled particles" OR "active nematic" OR "active gels"
The combination of these search terms was confirmed by three faculty members in this field to be able to capture the majority publications of active matter. Document types were limited to original research articles, review articles and early access articles. A total of 3,561 records were downloaded as “plain text files” with the option of “full record and cited references” in batches of 500.

CiteSpace Advanced version 6.2.R2 was used for data analysis. Data downloaded from Web of Science was processed in CiteSpace to remove duplicates before co-citation analysis. Date range was selected from January 2013 to February 2023 with time slicing of one year to analyze evolution of research trends in this field during the past ten years. Cited reference looking back time was unlimited, which meant there was no time limit to when the citation took place. As a selection criteria, the value of k was set up as 25, which is a modified g-index (Egghe, 2006). The smaller the k value, the fewer publications will be included in data analysis, which will produce fewer nodes in the visualization graph. However, this smaller data set will include the most influential publications in the research field under study.

The importance of the top four pivotal publications identified in this study were evaluated by two faculty members in this field independently through email interviews. Questions were asked as follows with an explanation of “pivotal publications” identified in CiteSpace:
“Could you write 2-3 sentences for each of the four articles about whether you think these publications were important; if they were important, what roles they played in the ‘active matter’ field?”

Results and Discussion

Publication Output and Research Disciplines

Results in this section were from the WOS Core Collection. Figure 3 displayed the number of publications of each year and the cumulative number of publications from 1990 to 2022 after duplicate removal. The line for number of publications of each year is typically a zig-zag line since each year’s publication number could fluctuate a lot. However, the cumulative number of publications can produce a smooth line which can be given a fitting equation in Excel. The cumulative publication number line and its fitting equation can tell us what rule the cumulative number of publications in a field follows: linear or exponential. On the other hand, since the cumulative number of publications is much higher than each year’s publication number, a secondary y-axis needs to be set up. Otherwise, the trend of every year’s publication number line would be indiscernible.

The duplicate removal report generated by CiteSpace lists the total number of publications along with the number of publications of each document type and each year. In addition to removing duplicated records, this process will also produce output files for each publication year, which will make subsequent data processing much faster. Since no duplicates were removed for this dataset (mainly based on DOIs), publication output generated by WOS or CiteSpace were identical in this current study. If duplicates were removed in the data pre-processing procedure, publication output produced by CiteSpace is more accurate. It can be seen from Figure 3 that there was a minor jump in the number of publications since 2004 and then a sharp increase since 2010. The increase in the number of publications in this research area corresponded to growth of interest in this field during the past ten years.

Web of Science Categories	Number of Publications	Percentage of total publications (3,561)
Physics Multidisciplinary	885	24.9
Chemistry Physical	536	15.1
Physics Mathematical	495	13.9
Materials Science Multidisciplinary	461	12.9
Physics Fluids Plasmas	397	11.1
Multidisciplinary Sciences	314	8.8
Polymer Science	259	7.3
Chemistry Multidisciplinary	231	6.5
Mechanics	220	6.2
Physics Applied	170	4.8

Table 1 displays the top ten disciplines in WOS based on number of publications related to “active matter” generated by WOS. “Physics” and its related disciplines (physics multidisciplinary, physics mathematical, physics fluids plasmas, and physics applied) ranked the first, third, 5th and 10th, respectively. “Chemistry” and its related disciplines (chemistry physical, polymer science, and chemistry multidisciplinary) ranked the second, 7th and 8th among all disciplines, respectively. The top ten WOS disciplines indicated that active matter research does not only focus on physics, but also covers chemistry and material science areas, which suggested it is multidisciplinary.

Analysis of Co-citation Research Clusters

The rest of the data in this study were all generated by CiteSpace. Nine major co-citation clusters were identified from 3,561 publications that comprised 81,896 distinct cited references of all document types. A total of 857 nodes and 6,182 links were formed. Visualization of these thematic clusters will tell us major research topics in a field, and how these topics developed with time. The average time of each cluster and cluster link walk-through will demonstrate what topics are research fronts. Typically, younger clusters are considered as research fronts. Cluster names generated by CiteSpace with citing article titles were displayed in Figure 4 with the tree-ring node type. The tree-rings can tell the history of citations for an article: the color represents when the citation happened based on the color legend (from purple to yellow with time progress), and the thickness of the ring corresponds to the number of citations received each year (Chen, 2006). The colors of each cluster were default colors from the software, which can be customized. Due to the size of Figure 4, the tree-rings can be seen better in large size nodes. The number of publications, silhouette score (S score), mean publication year and top terms (Log-Likelihood Ratio (LLR) Test method) of each cluster were listed in Supplementary Table S1. The S score measures the quality of cluster analysis results and ranges from -1 to 1, and a number larger than 0.7 indicates a precise clustering configuration (Rousseeuw, 1987). In Table S1, cluster S scores ranged from 0.813 to 0.968, which validated the classification results.

A link walkthrough of clusters can identify research trends and shifts over time (see video uploaded to OSF, no audio). Starting in 2013, there were five main research trends: clusters #0, #1 #2, #5, and #6. Cluster #0 was labeled as “active nematics” by titles of citing articles (LLR method). The major citing article of this cluster was a review article on various mechanisms of the man-made nanoswimmers and their applications in nonequilibrium physics (Bechinger et al., 2016). Cluster #1 was labeled as “self-propelled particle” by titles, and the major citing article in this cluster was a literature review on the role topology played in active matter and its applications in new materials (Shankar et al., 2022). Cluster #2 was labeled as “topological defect” by titles and it shared the same major citing article as cluster #1. Cluster #5 was labeled as “active motor” by titles and its major citing article was a perspective on chiral active matter (Liebchen & Levis, 2022). Cluster #6 was labeled as “bacterial suspension” by titles, featuring a major citing article which was a review on the active materials with life-like characteristics (Needleman & Dogic, 2017).

In 2014, clusters #0, #1 #2 and #5 remained research hotspots with some visible development into clusters #4 (entropy production), #7 (active gel) and #8 (periodic substrate). In 2015, clusters #0, #1, #5 and #6 were hot research areas. In 2016, active matter research witnessed new developments in clusters #3 and #4. Cluster #3 was labeled as “motility-induced phase separation” with a major citing article that summarized the equilibrium and nonequilibrium mechanics in active matter (O’Byrne et al., 2022). Cluster #4 was labeled as “entropy production,” featuring a major citing article that studied the equilibrium features in active systems (Dabelow et al., 2021). In 2017, research in active matter further evolved into clusters #4 and #7. Cluster #7 was labeled as “active gel” by titles, highlighted by a major citing article which was a review of actin–myosin systems in active soft matter (Alvarado et al., 2017). In 2018, new developments were visible in cluster #8, though clusters #0, #2, and #5 remained research hotspots. Cluster #8 was labeled as “periodic substrate” by titles, with a major citing article on the effect of quenched random potentials and torques on scalar active matter (Ro et al., 2021). In 2019, new development in the “active matter” field concentrated on clusters #2 (topological defect) and #4 (entropy production). Throughout the last three years (2020, 2021 and 2022), clusters #2 (topological defect), #3 (motility-induced phase separation), #4 (entropy production) and #5 (active motor) stayed as research hotspots.

The mean year of each cluster is another indicator of shift and progress in research interests (Supplementary Table S1). Cluster #6 (bacterial suspension) was the initial research hotspot with a mean year of 2005. The most recent research interests focused on clusters #2 (topological defect), #3 (motility-induced phase separation), #4 (entropy production) and #8 (periodic substrate), all with a mean year of 2013.

As demonstrated by the supplementary video, a link walkthrough between clusters could illustrate the shifts and evolution of research focus/ideas over time and research fronts. Compared with the classification functions in similar knowledge visualization tools, a unique feature of CiteSpace is the automatic labeling of clusters by title, keywords, or abstracts of citing articles. Under each cluster labeling option, there are three ways to select from: Latent Semantic Indexing (LSI), Log-Likelihood Ratio (LLR) Test, and Mutual Information (MI) (Chen et al., 2010). However, caution should be taken for older articles which were not indexed with keywords and abstracts, such as articles before 1991 in the WOS Core Collection. In that situation, cluster labeling by title is recommended, and labeling by title with the LLR method is also the default labeling method in CiteSpace. Cluster name displayed in graph is the top one term identified by each method, but it may not be the best scientific name. It has always been a challenge to summarize the themes of clusters in a visualized network. Although not perfect, the automatically generated cluster labels provided a starting point when expert input is not available. A new development in the 2023 version (6.R1) was the function to allow customized cluster labels. To relabel clusters, users can set up a TSV (tab-separated values) file by selecting the cluster names from the top terms listed under “cluster explorer.” The “cluster explorer” also lists major citing articles and cited references (knowledge base) of each cluster (see Figure 5 below). However, expertise knowledge would be required to precisely label clusters based on the themes of citing articles/cited references.

Mapping of research themes/interest and their shift over time have also been demonstrated in other case studies on the use of science mapping/bibliometric tools in library research support. For example, in one of the case studies with VOSviewer, co-keyword maps were used to track the progress of research in the field of “Internet of Things” (IoT) over time (MacDonald & Dressler, 2018). In another case study with VOSviewer, keyword overlay maps have also been applied to capture the evolvement of research interest in a specific institution (McAllister et al., 2022). Although keywords can tell us general information on what the research was about, the information revealed can be very limited. For example, in the IoT case study, for the 2014-2016 co-keywords map, the most visible co-keywords appeared to be IoT, “wireless sensor networks,” “security” and others. However, it was not very easy to discern what many keywords (such as management, model, big data, service) told us without context. Similarly, in the case study with VOSview to map changes of disciplines (McAllister et al., 2022), keywords clusters were formed to identify research hot spots in physical sciences and engineering of an institution. Meanings of keywords such as “water quality,” “agriculture,” and “fabrication” can be too broad in the area of membrane, microscopy and water treatment. Although keywords can tell us some information about a research area, they provide an incomplete picture, especially for librarians who are not experts of a specific field of study. On the other hand, in CiteSpace, a list of the themes of the co-citation network are extracted from titles, keywords or abstracts, instead of single keywords. What is more important, as demonstrated in Figure 5(a), for each cluster, major citing articles were listed, most with clickable DOI links. The first column “coverage” stands for how many cited references from that cluster were cited by that citing article. In Figure 5(b), cited references for the largest cluster were listed with DOIs and can be sorted based on centrality or burst strength scores. These cited references formed the knowledge base of the subject under study. With such information, it is very helpful to learn what this research topic/cluster was precisely about. In summary, CiteSpace provided a more detailed description of the evolvement of a research field with time and more information to detect the themes of research topics/fronts.

Most Cited Publications

Network citations	WOS citations*	Author(s) and year	Source title	Article title	Publication type
897	2,354	(Marchetti et al., 2013)	Reviews of modern physics	Hydrodynamics of soft active matter	review
532	4,725	(Vicsek et al., 1995)	Physical review letters	Novel type of phase transition in a system of self-driven particles	research
470	1,501	(Bechinger et al., 2016)	Reviews of modern physics	Active particles in complex and crowded environments	review
390	1,168	(Palacci et al., 2013)	Science	Living crystals of light-activated colloidal surfers	research
381	1,751	(Vicsek & Zafeiris, 2012)	Physics reports	Collective motion	review
377	985	(Ramaswamy, 2010)	Annual review of condensed matter physics	The mechanics and statistics of active matter	review
375	906	(Sanchez et al., 2012)	Nature	Spontaneous motion in hierarchically assembled active matter	research
347	784	(Cates & Tailleur, 2015)	Annual review of condensed matter physics	Motility-induced phase separation	review
319	712	(Fily & Marchetti, 2012)	Physical review letters	Athermal phase separation of self-propelled particles with no alignment	research
300	777	(Buttinoni et al., 2013)	Physical review letters	Dynamical clustering and phase separation in suspensions of self-propelled colloidal particles	research

* Citation counts collected on May 18, 2023

Citation counts or tracking have always been the traditional way to measure how influential an article has been since the citation link idea was introduced by Dr. Eugene Garfield (Garfield, 1955), and this metric is provided in many databases, such as WOS, Scopus and Google Scholar. Science mapping can be conducted at global or local levels or through a combination of the two (Chen & Song, 2019). Global citation counts can be retrieved for an article from the above three multidisciplinary databases and are viewed as the indicators of its influence in all scientific disciplines, which is referred to as globalism. On the other hand, localism focuses on the intellectual structure of a specific field. Bibliographic data in a local scope study are typically downloaded from a citation-based database with a query-based search. One feature of CiteSpace was to provide the number of local citations, which reflects the impact of an article in the field under study. Table 2 listed the most cited publications in the co-citation network with sources, article titles and publication types. It was noteworthy that the local citation counts in the network were lower than those in the literature, also known as global citation counts. This disparity in citation counts can be explained by the fact that the dataset analyzed in this case study was only a small subset from the WOS Core Collection. Citation orders in the two sets of citation counts could be different too. For example, in the co-citation network, the review article by Marchetti et al (2013) had higher citation counts than the research article by Vicsek et al (1995) (897 vs. 532), whereas in WOS Core Collection, the latter exhibited a higher citation count (4725 vs. 2354). This observation indicates that in the field of active matter, the review article by Marchetti et al (2013) seemingly received more attention than the research article by Vicsek et al (1995). The ratio of local to global citation can also be an indicator of how an article has been cited in different areas of science. For example, the review article by Marchetti et al (2013) has a ratio of local to global citation of 0.38 (897/2354), and the research article by Vicsek et al (1995) has such a ratio of 0.11 (532/4725). This difference suggests that the work by Vicsek et al (1995) was cited more often in fields other than active matter. Out of the top ten most cited publications, half were original research articles and half literature reviews. If focusing on research articles, the top three most cited articles were about a novel model of active matter (Vicsek et al., 1995), “living crystals” activated by light (Palacci et al., 2013), and active matter with biomimetic properties (Sanchez et al., 2012). In summary, CiteSpace can provide local citation counts, which reveals how influential an article is in the field under study. The articles that received top local citation counts and the journals they appeared in can be used complementarily with global citation data from major bibliographic databases for both research impact evaluation and collection development decisions.

Cited references variants can alter citation counts of a cited reference, as demonstrated in the Cited Reference Search in WOS (Clarivate, 2023). Based on the author’s experience of handling larger data sets with older records, books, book chapters, reference works (handbooks and encyclopedias), and older journal articles without DOIs assigned at the time of publication could appear as multiple nodes in the knowledge base (cite references). Due to inaccurate citations, CiteSpace could not merge all citation variants of the same cited reference. In this situation, manual inspection is required and nodes that belong to the same cited reference can be merged by setting up primary and secondary alias and then rerunning the visualization process.

Citation Bursts

The purpose of burst detection is to identify active research areas and emerging research trends. A citation burst means a surge of citation counts, which indicates the publication has attracted great interest from a research field and thus is an influential paper (Kleinberg, 2002). A citation burst demonstrates two properties: strength and duration. The number of citation bursts detected can be adjusted with parameters such as sensitivity of detection and minimum duration of years. A total of 31 citation bursts were detected with the current dataset when the minimum burst duration was set as five years (Supplementary Figure S1(a)).

Network citation	WOS citations*	Author(s) and year	Source title	Article title	Publication type
50	233	(Schwarz-Linek et al., 2012)	PNAS	Phase separation and rotor self-assembly in active particle suspensions	research
50	333	(Szabó et al., 2006)	Physical review E	Phase transition in the collective migration of tissue cells: Experiment and model	research
32	82	(Fielding et al., 2011)	Physical review E	Nonlinear dynamics and rheology of active fluids: Simulations in two dimensions	research
86	480	(Jiang et al., 2010)	Physical reviews letters	Active motion of a janus particle by self-thermophoresis in a defocused laser beam	research
28	220	(Czirók et al., 1997)	Journal of physics A: Mathematical and General	Spontaneously ordered motion of self-propelled particles	research
107	260	(Henkes et al., 2011)	Physical review E	Active jamming: Self-propelled soft particles at high density	research
45	200	(Ghosh et al., 2013)	Physical reviews letters	Self-propelled janus particles in a ratchet: numerical simulations	research
52	253	(Baskaran & Marchetti, 2009)	PNAS	Statistical mechanics and hydrodynamics of bacterial suspensions	research
52	206	(Cates et al., 2010)	PNAS	Arrested phase separation in reproducing bacteria creates a generic route to pattern formation	research
86	1329	(Couzin et al., 2002)	Journal of theoretical biology	Collective memory and spatial sorting in animal groups	research

* Citation counts collected on August 22, 2023

Table 3 displays the top ten publications with the strongest citation bursts that lasted at least five years. These ten publications were all research articles, with the top ranked publication being about the self-assembly of micro-motor active swimmers (Schwarz-Linek et al., 2012). The second one on the list was about a flocking model of tissue cells (Szabó et al., 2006), followed by a publication on the model of active fluids in two dimensions as the third on the list (Fielding et al., 2011).

If sorted by length of duration (Supplementary Figure S1(b)), the publication with the longest duration was seven years (2013-2019), which was a model on how groups of animals made decisions (Couzin et al., 2005). Five other publications followed on the list with a six-year burst duration.

When sorted by starting time, two publications stood out that indicated the most recent active research areas: one linked active nematic liquid crystals with the action mechanism of epithelia through topological defects (Saw et al., 2017), and the other explored how spindle-shaped cells controlled annihilation through topological defects (Duclos et al., 2017). These two most recent citation bursts suggested that topological defects of active matter in biological systems is a current active research area.

It was noteworthy that some citation burst publications exhibited relatively low network citations, such as the one by Fielding et al. (2011, 32 network citations) and the other by Czirók et al. (1997, 28 network citations). This reason is that burst detection is measured by a frequency function and these burst articles exhibited a citation surge in a short time compared with articles published at the same time. Very high citation counts do not necessarily lead to a citation burst (Chen et al., 2010).

Top ten most cited articles in the network (Table 2) and the top ten publications with the strongest citation bursts (Table 3) exhibited some distinct features. Most cited articles included both review and research articles, whereas publications with strongest citation bursts were all research articles. This difference indicated that citation counts for review articles appeared to be accumulated gradually over time. On the other hand, influential research articles in active research areas tended to receive relative high citations in a shorter period, such as a few years. It is also noteworthy to point out that Tables 2 and 3 included different source titles. For example, in Table 2, Annual Review of Condensed Matter Physics appeared twice, Physical Review Letters appeared three times, and Nature and Science each appeared once. On the other hand, in Table 3, PNAS were present three times, Physical Review E emerged three times, and Physical Review Letters appeared twice. In both tables, source titles included both physics journals and multidisciplinary scholarly journals. This information could be helpful in revealing important journals in a field for collection development purpose.

Burst detection is a function that is available in a few bibliometric/science mapping tools, based on a comparison of their features (Moral-Muñoz et al., 2020). Moral-Muñoz et al. (2020) listed three Science mapping analysis tools that provided this function: BiblioShiny, CiteSpace and SCI2. However, it seemed that the case study on the application of SCI2 to provide research support and collection development information did not discuss this function (Lewis & Alpi, 2017). On the other hand, the study that used BiblioShiny to analyze the application of artificial intelligence in libraries did not mention the burst detection function either (Borgohain et al., 2024). As a result, a comparison of burst detection could not be made among these tools in this current study. To summarize, citation burst detection provides another indicator to identify influential publications before they had time to receive enough citations to attract attention, and to reveal active research areas. Journal titles where these citation burst articles were published could also be considered in collection development decisions.

Pivotal Publications

Centrality	Network citations	WOS Citations	Author(s) and year	Source title	Article title	Publication type
0.11	897	2,354	(Marchetti et al., 2013)	Reviews of modern physics	Hydrodynamics of soft active matter	review
0.07	259	641*	(Wensink et al., 2012)	PNAS	Meso-scale turbulence in living fluids	research
0.06	319	712	(Fily & Marchetti, 2012)	Physical review letters	Athermal phase separation of self-propelled particles with no alignment	research
0.06	112	361*	(Cates, 2012)	Reports on Progress in Physics	Diffusive transport without detailed balance in motile bacteria: does microbiology need statistical physics?	review

*Citation counts collected on August 23, 2023

A vital and unique feature of CiteSpace is the identification of pivotal publications through betweenness centrality scores (Chen, 2004). Betweenness centrality was first introduced by Freeman in 1977 to measure social relationships (Freeman, 1977). The betweenness centrality score of a node is the shortest path between all pairs of nodes in a cluster through this node and it is calculated for every node in CiteSpace (Chen, 2005). A pivotal node is one with a high centrality score, which can appear between two clusters or inside a cluster. It serves as a gateway between clusters or occupies an influential place in a cluster, as demonstrated in Figure 2 and the nodes with a purple outline in Figure 4. In other words, links of cited references in different clusters must go through the pivot nodes, and these nodes are connectors in the network. A centrality score is normalized to [0,1] and tends to be a small number since it is a logarithm value. CiteSpace has been able to identify influential publications in research fields that the user had no subject knowledge with. For example, in 2012 two publications by Professor Yamanaka’s research team (Takahashi & Yamanaka, 2006; Takahashi et al., 2007) were recognized as transformative papers in regenerative medicine (Chen et al., 2012). A few months after this analysis, Professor Yamanaka was awarded the Nobel Prize in Physiology or Medicine (The Nobel Prize, n.d.).

The ability to identify pivotal publications can be useful to librarians when providing research impact evaluation, since this information is hard to obtain without subject knowledge. Four publications with the top centrality scores were listed in Table 4 and two of them also appeared on the most cited articles list. The publication with the highest centrality score was a review article on the development of theory behind active matter (Marchetti et al., 2013), followed by research on the statistics of self-sustained turbulence in bacterial suspensions (Wensink et al., 2012). The third publication on the list was a study on the athermal phase separation as a general property of active systems (Fily and Marchetti, 2012), and the fourth was a key issues review on whether statistical physics could be useful in microbiology (Cates, 2012). In summary, two of the four articles were literature reviews and two research articles. Three of the four source titles were physics journals, with the fourth PNAS as a multidisciplinary scholarly journal.

Author(s) and year	Expert A	Expert B
(Marchetti et al., 2013)	“This is one of the most important reviews in the field of active matter. It summarizes key methodologies in active matter from the perspective of continuum theory, as well as major discoveries in the field along this line. The models and theories presented in the review still dominate the research in active matter today.”	“It’s a review paper, a bit outdated by now; co-authored by early pioneers in the field, such as Sriram Ramaswamy and J Prost, people cite it whenever their paper deals with active matter thus the ‘impact’ is high. However, the true impact happened earlier, in the original papers by Ramaswamy, Prost, and others.”
(Wensink et al., 2012)	“This is also an important paper in active matter. It is one of the early works studying the statistics of active turbulence. The minimal continuum theory proposed here for active turbulence is still commonly used to describe the dynamics of active fluid, due to its simplicity.”	“A real groundbreaking paper with experiment and theory that show a non-inertial turbulent regime in active matter. An important step was in finding the spectrum of kinetic energy, Fig.4, which shows the difference between normal turbulence at high Reynolds numbers and turbulence at low Reynolds in active matter.”
(Fily and Marchetti, 2012)	“This paper represents a major breakthrough in the collective motion of self-propelled particles. It demonstrates that even without breaking rotational symmetry, a system of SPP can exhibit nontrivial properties such as nonequilibrium phase separation, giant number fluctuations, etc.”	“This is a paper rather limited in scope, as it explores active disks that show no orientational order. Usually active matter does show an orientational order since moving objects are elongated (fish, swimming bacteria, birds, sheep, etc.) The predicted result (phase separation) is not surprising, the only achievement being that this separation does not require an orientational order.”
(Cates, 2012)	“This is a very comprehensive review on the physics of swimming bacteria, from the statistics of random swimming trajectory, to various types of collective motion at different scenarios, highlighting the active nature of motile bacteria.”	“It is a kind of philosophical essay on whether the active matter systems such as swimming bacteria could be described by physics approaches developed for equilibrium systems. There is no clear cut answer as each situation requires a careful modeling whether statistical physics might help or might hurt.”

Table 5 listed the comments on the importance of these four articles from two faculty members in this field. Expert A is a junior physics faculty member who finished postdoctoral training at the lab of one of the main players in this field, with a h-index of 7 in Google Scholar. Expert B is an established materials science professor with a h-index of 79 in Google Scholar. Based on their comments, both experts agreed that the first article was an important review and cited by all active matter publications, although one expert pointed out it was published in 2013 and thus a little bit “outdated” probably because it was published ten years ago. Both experts pointed out the second publication was groundbreaking work, and its influence was still evident today. However, opinions on the third and fourth articles differed. Expert A deemed article three as a major breakthrough, while expert B viewed its influence as “limited.” Expert A believed article four was a “comprehensive” review, while expert B treated it as an “philosophical essay.” Based on comments from two experts, CiteSpace was able to identify pivotal publications in this field, such as the top two articles with high centrality scores. As far as how each expert views whether a publication is important in a field, the author discussed this question with Professor Rolfe Petschek (Emeritus Professor of Physics, Case Western Reserve University). Professor Petschek pointed out that individual expert’ opinions on whether a publication was influential could vary, and sometimes similar articles on the same topic could be published at about the same time (R. G. Petschek, personal communication, October 12, 2023). Another factor that led to mixed feedback on the importance of the top four articles could be that two of them were review articles, which typically contained multiple topics of a field. “Scientists often cite good, recent reviews rather than or in addition to the several original works on which they are based,” when a field depends on significant developments in several papers or the terminology has changed since the inception of the idea (R. G. Petschek, personal communication, November 21, 2024). In other words, real pivotal publications may be obscured by these highly cited reviews. As Expert B pointed out, if everybody else in this field has cited this review paper, then it must be important and hence everybody should cite it, even if this review article was off to the side of their underlying research areas. It is also necessary to point out in the current study, the most cited article happened to be the one with the highest centrality score. However, pivotal articles may not be the top highly cited articles. In other words, citation burst detection and betweenness centrality scores are both independent of citation counts. This feature can help us identify influential articles that have not received high citation counts (Chen, 2004). Active matter has only been an active field in the past twenty years and 3,561 publications were analyzed for co-citation network in this study. More publications will appear in this emerging area, producing a more connected co-citation network, which will provide a more comprehensive picture of pivotal publications in this area. However, bigger networks with more publications did not always lead to higher centrality scores, whose values depended mostly on the structure of the synthesized co-citation network (Chen, 2017).

The limited number of publications with top centrality scores evaluated by experts could also lead to mixed feedback from experts. The author was only able to ask the experts to evaluate the top four publications with high centrality scores. In the supplementary video, the third column of the table left to the visualization graph listed six publications with a centrality score of 0.05 and ten more with a centrality score of 0.04. It would be ideal if a rubric could be designed that the experts could then use to evaluate the potentially pivotal nature of the publications beyond the top four, which could reflect more accurately the robustness of their evaluation and provide a more comprehensive picture of influential papers in this field. In addition, review articles and original research articles can be evaluated separately, since review articles typically discuss several topics and can be cited for each of the topics they mentioned, or review articles could be cited just because everybody else is citing it. On the contrary, original research articles have a much narrower focus and readers can typically easily tell why they were cited. Furthermore, a combined evaluation of both top centrality publications and citation burst publications could produce a more complete picture of influential articles.

Identification of pivotal publications is a feature that seems unavailable in other major bibliometric/science mapping tools. The use of this feature can provide unique information when librarians provide research impact evaluation service, since recent influential publications that have not reached high citation threshold can be recognized through this feature.

Challenges

The learning curve of using CiteSpace includes pre-processing of data, understanding various functions (types of analysis) and the parameters associated with these functions, and, most importantly, data interpretation. The software can require a high computation power especially when a large number of publications are processed. There is a new function of labeling thematic clusters based on publication titles with Open AI. However, a payment to Open AI is required. The author received “rate limit reached” message after a small payment to Open AI. Citations to books and book chapters can be obscure. In this situation, articles that have cited these books or book chapters must be tracked down in order to find the bibliographic information for these cited books or book chapters.

Conclusions and Limitations

In this study, the concept “active matter” was used as an example to demonstrate how to use co-citation analysis to identify research hotspots and influential publications. Articles of active matter from the past ten years were analyzed for co-citation networks to group and label research topics. Recent research hotpots were identified by both the mean year of topic clusters and citation bursts. Pivotal publications were identified based on betweenness centrality, and comments from two experts confirmed the importance of the top two pivotal publications. Comparisons of visualization results from CiteSpace and other frequently used bibliometric/science mapping tools were also presented. Techniques illustrated in this study can be applied in other research fields to provide research trend analysis and research impact evaluation services and for collection development purposes.

One limitation in the current study was that the author was only able to ask experts in the “active matter” area to evaluate the top four articles with high centrality scores, two of which were review articles. A rubric could have been designed for the experts to evaluate the potentially pivotal nature of more original research articles on the centrality score list and the citation burst list, which could produce a more complete picture of influential publications in this field.

Data used in this study were solely collected from WOS Core Collection, which is a relatively selective bibliographic database. However, this database does provide the most comprehensive citation data for bibliometric analysis. Data from WOS Core Collection is also the default data format for CiteSpace, and so no data will be lost since data conversion is not required. Data from other databases such as Scopus, PubMed and lens.org is converted to the WOS data format before data analysis in CiteSpace. Another drawback of collecting data from WOS is that the download limit is 500 records each time, which means data collection can be time consuming and tedious if a large data set is being studied.

Another limitation in this study is that citations could not account for the reasons for the counted citations: whether they were vital, incidental, critical, or gift citations. Other citation biases can include self-citations and citations from journals with high impact factors. Finally, another limitation when using bibliometrics indicators to identify innovative research is that novel research could suffer from delayed recognition (Wang et al., 2017).

Acknowledgements

The author is thankful to Professor Robin Selinger for discussion on active matter and for confirming search terms in WOS. The author would like to acknowledge Professor Zhihong You at the Physics Department of Xiamen University (Xiamen, China), for confirming search terms in WOS and for answering interview questions through email. The author is also very grateful to Professor Oleg Lavrentovich for confirming search terms in WOS and for answering interview questions via email. The author would like to thank Daniel Dotson, mathematical sciences librarian & science education specialist at the University Libraries of the Ohio State University, for reviewing this manuscript and providing valuable feedback. The author benefitted greatly from the advice of Professor Yin Zhang at the School of Information at Kent State University when revising this manuscript.

References

Alvarado, J., Sheinman, M., Sharma, A., MacKintosh, F. C., & Koenderink, G. H. (2017). Force percolation of contractile active gels. Soft Matter, 13(34), 5624-5644. https://doi.org/10.1039/C7SM00834A

Baskaran, A., & Marchetti, M. C. (2009). Statistical mechanics and hydrodynamics of bacterial suspensions. Proceedings of the National Academy of Sciences, 106(37), 15567-15572. https://doi.org/10.1073/pnas.0906586106

Bechinger, C., Di Leonardo, R., Löwen, H., Reichhardt, C., Volpe, G., & Volpe, G. (2016). Active particles in complex and crowded environments. Reviews of Modern Physics, 88(4), 045006. https://doi.org/10.1103/RevModPhys.88.045006

Borgohain, D. J., Bhardwaj, R. K., & Verma, M. K. (2024). Mapping the literature on the application of artificial intelligence in libraries (AAIL): A scientometric analysis. Library Hi Tech, 42(1), 149-179. https://doi.org/10.1108/LHT-07-2022-0331

Braam, R. R., Moed, H. F., & Van Raan, A. F. (1991). Mapping of science by combined co‐citation and word analysis. II: Dynamical aspects. Journal of the American Society for Information Science, 42(4), 252-266. https://doi.org/10.1002/(SICI)1097-4571(199105)42:4<252::AID-ASI2>3.0.CO;2-G

Buttinoni, I., Bialké, J., Kümmel, F., Löwen, H., Bechinger, C., & Speck, T. (2013). Dynamical clustering and phase separation in suspensions of self-propelled colloidal particles. Physical Review Letters, 110(23), 238301. https://doi.org/10.1103/PhysRevLett.110.238301

Cates, M. E. (2012). Diffusive transport without detailed balance in motile bacteria: Does microbiology need statistical physics? Reports on Progress in Physics, 75(4), 042601. https://doi.org/10.1088/0034-4885/75/4/042601

Cates, M. E., Marenduzzo, D., Pagonabarraga, I., & Tailleur, J. (2010). Arrested phase separation in reproducing bacteria creates a generic route to pattern formation. Proceedings of the National Academy of Sciences, 107(26), 11715-11720. https://doi.org/10.1073/pnas.1001994107

Cates, M. E., & Tailleur, J. (2015). Motility-induced phase separation. Annual Review of Condensed Matter Physics, 6(1), 219-244. https://doi.org/10.1146/annurev-conmatphys-031214-014710

Chen, C. (2004). Searching for intellectual turning points: Progressive knowledge domain visualization. Proceedings of the National Academy of Sciences, 101(Suppl 1), 5303-5310. https://doi.org/10.1073/pnas.0307513100

Chen, C. (2005, January 10-13). The centrality of pivotal points in the evolution of scientific networks [Paper presentation]. 10th International Conference on Intelligent User Interfaces, San Diego, CA, United States. https://doi.org/10.1145/1040830.1040859

Chen, C. (2006). CiteSpace II: Detecting and visualizing emerging trends and transient patterns in scientific literature. Journal of the American Society for Information Science and Technology, 57(3), 359-377. https://doi.org/10.1002/asi.20317

Chen, C. (2017). Science mapping: A systematic review of the literature. Journal of data and information science, 2(2), 1-40. https://doi.org/10.1515/jdis-2017-0006

Chen, C., Hu, Z., Liu, S., & Tseng, H. (2012). Emerging trends in regenerative medicine: A scientometric analysis in CiteSpace. Expert Opinion on Biological Therapy, 12(5), 593-608. https://doi.org/10.1517/14712598.2012.674507

Chen, C., Ibekwe‐SanJuan, F., & Hou, J. (2010). The structure and dynamics of cocitation clusters: A multiple‐perspective cocitation analysis. Journal of the American Society for Information Science and Technology, 61(7), 1386-1409. https://doi.org/10.1002/asi.21309

Chen, C., & Song, M. (2019). Visualizing a field of research: A methodology of systematic scientometric reviews. PloS One, 14(10), e0223994. https://doi.org/10.1371/journal.pone.0223994

Chen, Y., & Zhou, L. (2021). Library research support services in China’s universities of traditional medicine: Understanding user requirements. Health Information & Libraries Journal, 38(4), 313-318. https://doi.org/10.1111/hir.12411

Clarivate. (2023, November 12). Web of Science help: Cited references search. https://webofscience.help.clarivate.com/en-us/Content/cited-reference-search.htm

Corrall, S., Kennan, M. A., & Afzal, W. (2013). Bibliometrics and research data management services: Emerging trends in library support for research. Library Trends, 61(3), 636-674. https://doi.org/10.1353/lib.2013.0005

Couzin, I. D., Krause, J., Franks, N. R., & Levin, S. A. (2005). Effective leadership and decision-making in animal groups on the move. Nature, 433(7025), 513-516. https://doi.org/10.1038/nature03236

Couzin, I. D., Krause, J., James, R., Ruxton, G. D., & Franks, N. R. (2002). Collective memory and spatial sorting in animal groups. Journal of Theoretical Biology, 218(1), 1-11. https://doi.org/10.1006/jtbi.2002.3065

Czirók, A., Stanley, H. E., & Vicsek, T. (1997). Spontaneously ordered motion of self-propelled particles. Journal of Physics A: Mathematical and General, 30(5), 1375. https://doi.org/10.1088/0305-4470/30/5/009

Dabelow, L., Bo, S., & Eichhorn, R. (2021). How irreversible are steady-state trajectories of a trapped active particle? Journal of Statistical Mechanics: Theory and Experiment, 2021(3), 033216. https://doi.org/10.1088/1742-5468/abe6fd

Drummond, R. (2014). RIMS revisited: The evolution of the research impact measurement service at UNSW Library. Australian Academic & Research Libraries, 45(4), 309-322. https://doi.org/10.1080/00048623.2014.945065

Drummond, R., & Wartho, R. (2009). RIMS: The research impact measurement service at the University of New South Wales. Australian Academic & Research Libraries, 40(2), 76-87. https://doi.org/10.1080/00048623.2009.10721387

Duclos, G., Erlenkämper, C., Joanny, J.-F., & Silberzan, P. (2017). Topological defects in confined populations of spindle-shaped cells. Nature Physics, 13(1), 58-62. https://doi.org/10.1038/nphys3876

Egghe, L. (2006). Theory and practise of the g-index. Scientometrics, 69(1), 131-152. https://doi.org/10.1007/s11192-006-0144-7

Fernández-Marcial, V., Costa, L. M., & González-Solar, L. (2016, June 5-9). Top universities, top libraries: Do research services in academic libraries contribute to university output [Paper Presentation]? International Association of Scientific and Technological University Libraries (IATUL) Conference, Halifax, NS, Canada. https://docs.lib.purdue.edu/iatul/2016/spaces/2

Fielding, S. M., Marenduzzo, D., & Cates, M. E. (2011). Nonlinear dynamics and rheology of active fluids: Simulations in two dimensions. Physical Review E, 83(4), 041910. https://doi.org/10.1103/PhysRevE.83.041910

Fily, Y., & Marchetti, M. C. (2012). Athermal phase separation of self-propelled particles with no alignment. Physical Review Letters, 108(23), 235702. https://doi.org/10.1103/PhysRevLett.108.235702

Freeman, L. C. (1977). A set of measures of centrality based on betweenness. Sociometry, 40(1), 35-41. https://doi.org/10.2307/3033543

Garfield, E. (1955). Citation indexes for science: A new dimension in documentation through association of ideas. Science, 122(3159), 108-111. https://doi.org/10.1126/science.122.3159.108

Garfield, E., Pudovkin, A. I., & Istomin, V. S. (2003). Why do we need algorithmic historiography? Journal of the American Society for Information Science and Technology, 54(5), 400-412. https://doi.org/10.1002/asi.10226

Ghosh, P. K., Misko, V. R., Marchesoni, F., & Nori, F. (2013). Self-propelled Janus particles in a ratchet: Numerical simulations. Physical Review Letters, 110(26), 268301. https://doi.org/10.1103/PhysRevLett.110.268301

Hallatschek, O., Datta, S. S., Drescher, K., Dunkel, J., Elgeti, J., Waclaw, B., & Wingreen, N. S. (2023). Proliferating active matter. Nature Reviews Physics, 5(7), 407-419. https://doi.org/10.1038/s42254-023-00593-0

Henkes, S., Fily, Y., & Marchetti, M. C. (2011). Active jamming: Self-propelled soft particles at high density. Physical Review E, 84(4), 040301. https://doi.org/10.1103/PhysRevE.84.040301

Hou, J., Yang, X., & Chen, C. (2018). Emerging trends and new developments in information science: A document co-citation analysis (2009–2016). Scientometrics, 115, 869-892. https://doi.org/10.1007/s11192-018-2695-9

Howie, J., & Kara, H. (2022). Research support in New Zealand university libraries. New Review of Academic Librarianship, 28(1), 7-36. https://doi.org/10.1080/13614533.2019.1700535

Jiang, H. R., Yoshinaga, N., & Sano, M. (2010). Active motion of a Janus particle by self-thermophoresis in a defocused laser beam. Physical Review Letters, 105(26), 268302. https://doi.org/10.1103/PhysRevLett.105.268302

Kleinberg, J. (2002, July 23-26). Bursty and hierarchical structure in streams [Paper presentation]. ACM SIGKDD International Conference on Knowledge Discovery and Data Mining. Edmonton, AB, Canada. https://doi.org/10.1145/775047.775061

Lewis, D. M., & Alpi, K. M. (2017). Bibliometric network analysis and visualization for serials librarians: An introduction to Sci2. Serials Review, 43(3-4), 239-245. https://doi.org/10.1080/00987913.2017.1368057

Liebchen, B., & Levis, D. (2022). Chiral active matter. Europhysics Letters, 139(6), 67001. https://doi.org/10.1209/0295-5075/ac8f69

Lin, S., Shen, T., & Guo, W. (2021). Evolution and emerging trends of Kansei Engineering: A visual analysis based on CiteSpace. IEEE Access, 9, 111181-111202. https://doi.org/10.1109/ACCESS.2021.3102606

MacDonald, K. I., & Dressler, V. (2018). Using citation analysis to identify research fronts: A case study with the Internet of Things. Science & Technology Libraries, 37(2), 171-186. https://doi.org/10.1080/0194262X.2017.1415183

Marchetti, M. C., Joanny, J. F., Ramaswamy, S., Liverpool, T. B., Prost, J., Rao, M., & Simha, R. A. (2013). Hydrodynamics of soft active matter. Reviews of Modern Physics, 85(3), 1143. https://doi.org/10.1103/RevModPhys.85.1143

McAllister, J. T., Lennertz, L., & Atencio Mojica, Z. (2022). Mapping a discipline: A guide to using VOSviewer for bibliometric and visual analysis. Science & Technology Libraries, 41(3), 319-348. https://doi.org/10.1080/0194262X.2021.1991547

Miller, A., Johnson, J., Cline, L., Edgar, W., Fischer, S., & Jackson-Brown, G. (2023). Research services and spaces as expressions of the scholarly information seeking habits and behavior of university faculty. Journal of Library Administration, 63(5), 633-651. https://doi.org/10.1080/01930826.2023.2219597

Moral-Muñoz, J. A., Herrera-Viedma, E., Santisteban-Espejo, A., & Cobo, M. J. (2020). Software tools for conducting bibliometric analysis in science: An up-to-date review. Profesional de la información/Information Professional, 29(1). https://doi.org/10.3145/epi.2020.ene.03

Needleman, D., & Dogic, Z. (2017). Active matter at the interface between materials science and cell biology. Nature Reviews Materials, 2(9), 1-14. https://doi.org/10.1038/natrevmats.2017.48

The Nobel Prize. (n.d.). The Nobel Prize in Physiology or Medicine 2012. https://www.nobelprize.org/prizes/medicine/2012/summary/

O’Byrne, J., Kafri, Y., Tailleur, J., & van Wijland, F. (2022). Time irreversibility in active matter, from micro to macro. Nature Reviews Physics, 4(3), 167-183. https://doi.org/10.1038/s42254-021-00406-2

Palacci, J., Sacanna, S., Steinberg, A. P., Pine, D. J., & Chaikin, P. M. (2013). Living crystals of light-activated colloidal surfers. Science, 339(6122), 936-940. https://doi.org/10.1126/science.1230020

Pan, X., Yan, E., Cui, M., & Hua, W. (2018). Examining the usage, citation, and diffusion patterns of bibliometric mapping software: A comparative study of three tools. Journal of Informetrics, 12(2), 481-493. https://doi.org/10.1016/j.joi.2018.03.005

Persson, O. (1994). The intellectual base and research fronts of JASIS 1986–1990. Journal of the American Society for Information Science, 45(1), 31-38. https://doi.org/10.1002/(SICI)1097-4571(199401)45:1<31::AID-ASI4>3.0.CO;2-G

Ramaswamy, S. (2010). The mechanics and statistics of active matter. Annual Review of Condensed Matter Physics, 1(1), 323-345. https://doi.org/10.1146/annurev-conmatphys-070909-104101

Rawat, K. S., & Sood, S. K. (2021). Knowledge mapping of computer applications in education using CiteSpace. Computer Applications in Engineering Education, 29(5), 1324-1339. https://doi.org/10.1002/cae.22388

Ro, S., Kafri, Y., Kardar, M., & Tailleur, J. (2021). Disorder-induced long-ranged correlations in scalar active matter. Physical Review Letters, 126(4), 048003. https://doi.org/10.1103/PhysRevLett.126.048003

Rousseeuw, P. J. (1987). Silhouettes: A graphical aid to the interpretation and validation of cluster analysis. Journal of Computational and Applied Mathematics, 20, 53-65. https://doi.org/10.1016/0377-0427(87)90125-7

Sabe, M., Chen, C., Perez, N., Solmi, M., Mucci, A., Galderisi, S., Strauss, G. P., & Kaiser, S. (2023). Thirty years of research on negative symptoms of schizophrenia: A scientometric analysis of hotspots, bursts, and research trends. Neuroscience & Biobehavioral Reviews, 144, 104979. https://doi.org/10.1016/j.neubiorev.2022.104979

Sanchez, T., Chen, D. T., DeCamp, S. J., Heymann, M., & Dogic, Z. (2012). Spontaneous motion in hierarchically assembled active matter. Nature, 491(7424), 431-434. https://doi.org/10.1038/nature11591

Saw, T. B., Doostmohammadi, A., Nier, V., Kocgozlu, L., Thampi, S., Toyama, Y., Marcq, P., Lim, C. T., Yeomans, J. M., & Ladoux, B. (2017). Topological defects in epithelia govern cell death and extrusion. Nature, 544(7649), 212-216. https://doi.org/10.1038/nature21718

Schwarz-Linek, J., Valeriani, C., Cacciuto, A., Cates, M. E., Marenduzzo, D., Morozov, A. N., & Poon, W. C. K. (2012). Phase separation and rotor self-assembly in active particle suspensions. Proceedings of the National Academy of Sciences, 109(11), 4052-4057. https://doi.org/10.1073/pnas.1116334109

Shankar, S., Souslov, A., Bowick, M. J., Marchetti, M. C., & Vitelli, V. (2022). Topological active matter. Nature Reviews Physics, 4(6), 380-398. https://doi.org/10.1038/s42254-022-00445-3

Si, L., Zeng, Y., Guo, S., & Zhuang, X. (2019). Investigation and analysis of research support services in academic libraries. The Electronic Library, 37(2), 281-301. https://doi.org/10.1108/EL-06-2018-0125

Small, H. (1973). Co‐citation in the scientific literature: A new measure of the relationship between two documents. Journal of the American Society for Information Science, 24(4), 265-269. https://doi.org/10.1002/asi.4630240406

Small, H., & Griffith, B. C. (1974). The structure of scientific literatures I: Identifying and graphing specialties. Science Studies, 4(1), 17-40. https://doi.org/10.1177/030631277400400102

Small, H., & Sweeney, E. (1985). Clustering the Science Citation Index® using co-citations: I. A comparison of methods. Scientometrics, 7, 391-409. https://doi.org/10.1007/BF02017157

Sugi, T., Ito, H., & Nagai, K. H. (2021). Collective pattern formations of animals in active matter physics. Biophysics and Physicobiology, 18, 254-262. https://doi.org/10.2142/biophysico.bppb-v18.028

Szabo, B., Szöllösi, G. J., Gönci, B., Jurányi, Z., Selmeczi, D., & Vicsek, T. (2006). Phase transition in the collective migration of tissue cells: Experiment and model. Physical Review E, 74(6), 061908. https://doi.org/10.1103/PhysRevE.74.061908

Takahashi, K., Tanabe, K., Ohnuki, M., Narita, M., Ichisaka, T., Tomoda, K., & Yamanaka, S. (2007). Induction of pluripotent stem cells from adult human fibroblasts by defined factors. Cell, 131(5), 861-872. https://doi.org/10.1016/j.cell.2007.11.019

Takahashi, K., & Yamanaka, S. (2006). Induction of pluripotent stem cells from mouse embryonic and adult fibroblast cultures by defined factors. Cell, 126(4), 663-676. https://doi.org/10.1016/j.cell.2006.07.024

Van Eck, N., & Waltman, L. (2010). Software survey: VOSviewer, a computer program for bibliometric mapping. Scientometrics, 84(2), 523-538. https://doi.org/10.1007/s11192-009-0146-3

Vicsek, T., Czirók, A., Ben-Jacob, E., Cohen, I., & Shochet, O. (1995). Novel type of phase transition in a system of self-driven particles. Physical Review Letters, 75(6), 1226. https://doi.org/10.1103/PhysRevLett.75.1226

Vicsek, T., & Zafeiris, A. (2012). Collective motion. Physics Reports, 517(3-4), 71-140. https://doi.org/10.1016/j.physrep.2012.03.004

Wang, J., Veugelers, R., & Stephan, P. (2017). Bias against novelty in science: A cautionary tale for users of bibliometric indicators. Research Policy, 46(8), 1416-1436. https://doi.org/10.1016/j.respol.2017.06.006

Wang, Y., Mushtaq, R. T., Ahmed, A., Rehman, M., Khan, A. M., Sharma, S., Ishfaq, K., Ali, H., & Gueye, T. (2022). Additive manufacturing is sustainable technology: CiteSpace based bibliometric investigations of fused deposition modeling approach. Rapid Prototyping Journal, 28(4), 654-675. https://doi.org/10.1108/RPJ-05-2021-0112

Wei, Y., Zhang, Q., Guo, J., & Chen, M. (2023). Learning to teach through noticing: A bibliometric review of teacher noticing research in mathematics education during 2006–2021. Humanities and Social Sciences Communications, 10(1), 1-15. https://doi.org/10.1057/s41599-023-01718-7

Wensink, H. H., Dunkel, J., Heidenreich, S., Drescher, K., Goldstein, R. E., Löwen, H., & Yeomans, J. M. (2012). Meso-scale turbulence in living fluids. Proceedings of the National Academy of Sciences, 109(36), 14308-14313. https://doi.org/10.1073/pnas.1202032109

Wilson, T. (2017, June 18-22). Why is the library involved in research support? Demonstrating why and how the library can contribute and add value to the university research priorities [Paper presentation]. International Association of Scientific and Technological University Libraries (IATUL) Conference, Bolzano, BZ, Italy. https://docs.lib.purdue.edu/iatul/2017/research/3

Wu, J. F., & Tsai, H. L. (2022). An explication of HistCite: Updates, modifications, and a variety of applications. Serials Review, 48(1-2), 41-48. https://doi.org/10.1080/00987913.2022.2101821

Yang, J., Cheng, C., Shen, S., & Yang, S. (2017, March 10-12). Comparison of complex network analysis software: Citespace, SCI 2 and Gephi. Institute of Electrical and Electronics Engineers (IEEE) 2nd International Conference on Big data Analysis (ICBDA), Beijing, China. https://doi.org/10.1109/ICBDA.2017.8078800

Ye, L. (2019). Chinese academic library research evaluation services. Journal of Library Administration, 59(1), 97-128. https://doi.org/10.1080/01930826.2018.1549416

Zakaria, M. S. (2021). Data visualization as a research support service in academic libraries: An investigation of world-class universities. The Journal of Academic Librarianship, 47(5), 102397. https://doi.org/10.1016/j.acalib.2021.102397

Zhang, J., Quoquab, F., & Mohammad, J. (2023). Plastic and sustainability: A bibliometric analysis using VOSviewer and CiteSpace. Arab Gulf Journal of Scientific Research, 42(1), 44-67. https://doi.org/10.1108/AGJSR-10-2022-0225

Supplementary Material

The supplementary material including Table S1 of major clusters names and top terms in each cluser, Figure S2(a) of citation bursts sorted based on strength, and Figure S2(b) of citation bursts sorted based on duration of the bursts is available at https://journals.library.ualberta.ca/istl/index.php/istl/article/view/2779/2821

This work is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License.