PhD Seminar Series: Bibliometric Research Synthesis

bibliometrix: An R-tool for

comprehensive science mapping analysis
Massimo Aria and Corrado Cuccurullo
[email protected]; [email protected]

Bibliometrix package www.bibliometrix.org

Practice experience
• Seminars goals • Lab activities • Effective learning and usefulness
The aim of this seminar cycle is twofold. In this seminar, doctoral students have to synthesize a The exercise proves useful for all involved. Doctoral
large volume of studies and organize articles into students learn valuable skills in analysing and
1. First, we want to bring together in a single different intellectual and conceptual maps that conceptualizing vast amounts of literature. This is a
seminar cycle all the knowledge on research represents distinct research streams and useful skill for any researcher. Moreover, doctoral
synthesis through a recommended workflow, interrelationships. The assignment culminates with students will be able to use bibliometric maps to
from problem formulation to report writing. doctoral students presenting (a) graphical identify gaps and trends in the literature.
representations of their bibliometric analysis, (b) a
2. Second, we present our open-source bibliometrix depiction of how each stream relates to other Given the value of this experience, we ask doctoral
R-package for performing comprehensive streams, and (c ) a list of main contributing authors students to present and defend their own maps of
bibliometric analyses, and discuss how and/or works in each stream and substream. literature.
bibliometrix is a valid tool for performing
bibliometric studies. We illustrate the main
bibliometrix functions in the workflow, using
topics selected by the participants to the
seminars.
Key readings
• Research Synthesis • General science mapping workflow • Co-word analysis and Research Front
• Thomé, A. M. T., Scavarda, L. F., & Scavarda, A. J. (2016). • Aria, M. & Cuccurullo, C. (2017). bibliometrix: An R-tool for • Cuccurullo, C., Aria, M., & Sarto, F. (2016). Foundations and
Conducting systematic literature review in operations comprehensive science mapping analysis, Journal of trends in performance management. A twenty-five years
management. Production Planning & Control, 27(5), 408- Informetrics, 11(4), pp 959-975 bibliometric analysis in business and public administration
420. domains, Scientometrics, DOI: 10.1007/s11192-016-1948-8.
• Cooper, H. (2015). Research synthesis and meta-analysis: A • Cobo, M. J., Lopez-Herrera, A. G., Herrera-Viedma, E., & • Cobo, M. J., López-Herrera, A. G., Herrera-Viedma, E., &
step-by-step approach (Vol. 2). Sage publications. Herrera, F. (2011). Science Mapping Software Tools: Herrera, F. (2011). An approach for detecting, quantifying,
Review, Analysis, and Cooperative Study Among Tools. and visualizing the evolution of a research field: A practical
• Briner RB , Denyer D (2012) Systematic review and
Journal of the American Society for Information Science and application to the fuzzy sets theory field. Journal of
evidence synthesis as a practice and scholarship tool in
Technology. Informetrics, 5(1), 146-166.
Rousseau, D. M. (Ed.). (2012). The Oxford handbook of
evidence-based management. Oxford University Press.
• bibliometrix R-package (http://www.bibliometrix.org)
• Moher, D., Liberati, A., Tetzlaff, J., Altman, D. G., & Prisma
Group. (2009). Preferred reporting items for systematic • Bibliometrix Tutorial • Datascience and big data
reviews and meta-analyses: the PRISMA statement. PLoS
medicine, 6(7), e1000097. • Bibliometrix function map • George, G., Osinga, E. C., Lavie, D., & Scott, B. A. (2016). Big
data and data science methods for management
• Massaro, M., Dumay, J., & Guthrie, J. (2016). On the
shoulders of giants: undertaking a structured literature • Citation Indicators research. Academy of Management Journal, 59(5), 1493-
1507.
review in accounting. Accounting, Auditing & Accountability • Waltman, L. (2016). A review of the literature on citation
Journal, 29(5), 767-801. impact indicators. Journal of Informetrics, 10(2), 365-391. • Sivarajah, U., Kamal, M. M., Irani, Z., & Weerakkody, V.
(2017). Critical analysis of Big Data challenges and analytical
• Webster, J., & Watson, R. T. (2002). Analyzing the past to methods. Journal of Business Research, 70, 263-286.
prepare for the future: Writing a literature review. MIS • Intellectual Map
quarterly, xiii-xxiii. • Yang, S., Han, R., Wolfram, D., & Zhao, Y. (2016). Visualizing
• Torraco, R. J. (2005). Writing integrative literature reviews: the intellectual structure of information science (2006–
Guidelines and examples. Human resource development 2015): Introducing author keyword coupling analysis. Journal
review, 4(3), 356-367. of Informetrics, 10(1), 132-150.
Key readings in Strategy
• Conceptual Reviews, 10(1), 1-23. • Dagnino, G. B., Levanti, G., Minà, A., & Picone, P. M. (2015).
Interorganizational network and innovation: A bibliometric
• Nag, R., Hambrick, D. C., & Chen, M. J. (2007). What is • Phelan, S. E., Ferreira, M., & Salvador, R. (2002). The first study and proposed research agenda. Journal of Business &
strategic management, really? Inductive derivation of a twenty years of the Strategic Management Industrial Marketing, 30(3/4), 354-377.
consensus definition of the field. Strategic management Journal. Strategic Management Journal, 23(12), 1161-1168.
journal, 28(9), 935-955.
• Ronda‐Pupo, G. A., & Guerras‐Martin, L. Á. (2012).
• Hoskisson, R. E., Hitt, M. A., Wan, W. P., & Yiu, D. (1999). Dynamics of the evolution of the strategy concept 1962–
Theory and research in strategic management: Swings of a 2008: a co‐word analysis. Strategic Management
pendulum. Journal of management, 25(3), 417-456. Journal, 33(2), 162-188.

• Adcroft, A., & Willis, R. (2008). A snapshot of strategy • Maia, J. L., Serio, L. C., & Alves Filho, A. G. (2015). Almost
research 2002-2006. Journal of Management History, 14(4), two decades after: a bibliometric effort to map research on
313-333. strategy as practice using two data sources. European
Journal of Economics, Finance and Administrative
• Bibliometric articles (General) Sciences, 73, 7-31.

• Ramos‐Rodríguez, A. R., & Ruíz‐Navarro, J. (2004). Changes • Acedo, F. J., Barroso, C., & Galan, J. L. (2006). The
in the intellectual structure of strategic management resource‐based theory: dissemination and main
research: A bibliometric study of the Strategic Management trends. Strategic Management Journal, 27(7), 621-636.
Journal, 1980–2000. Strategic Management Journal, 25(10),
981-1004. • Vogel, R., & Güttel, W. H. (2013). The dynamic capability
view in strategic management: a bibliometric
• Nerur, S. P., Rasheed, A. A., & Natarajan, V. (2008). The review. International Journal of Management
intellectual structure of the strategic management field: An Reviews, 15(4), 426-446.
author co‐citation analysis. Strategic Management
Journal, 29(3), 319-336. • Di Stefano, G., Peteraf, M., & Verona, G. (2010). Dynamic
capabilities deconstructed‡: a bibliographic investigation
• Furrer, O., Thomas, H., & Goussevskaia, A. (2008). The into the origins, development, and future directions of the
structure and evolution of the strategic management field: research domain. Industrial and Corporate Change, 19(4),
A content analysis of 26 years of strategic management 1187-1204.
research. International Journal of Management
Context
• Topic Relevance • Bibliometrics
The number of academic publications is increasing at a Scholars use different qualitative and quantitative
rapid pace and it is becoming increasingly unfeasible to literature reviewing approaches to understand and
remain current with everything that is being published. organize earlier findings. Among these, bibliometrics
Moreover, the emphasis on empirical contributions has has the potential to introduce a systematic,
resulted in voluminous and fragmented research transparent, and reproducible review process based on
streams. This hampers the ability to accumulate the statistical measurement of science, scientists, or
knowledge and actively collect evidence through a set scientific activity. Unlike other techniques,
of previous research papers. Therefore, literature bibliometrics provides more objective and reliable
reviews are increasingly assuming a crucial role in analyses. The overwhelming volume of new Altmetrics
synthesizing past research findings to effectively use information, conceptual developments, and data are
the existing knowledge base, advance a line of the milieu where bibliometrics becomes useful by
research, and provide evidence-based insight into the providing a structured analysis to a large body of
practice of exercising and sustaining professional information, to infer trends over time, themes
judgment and expertise. researched, identify shifts in the boundaries of the
disciplines, to detect the most prolific scholars and
institutions, and to present the “big picture” of extant
research.

Bibliometrics for:
• Research valuation
• Science Mapping
Bibliometrix
• Complexity of bibliometric analysis • Bibliometrix: one tool for the whole bibliometric
workflow
Although over time, the use of bibliometrics has been
extended to all disciplines, bibliometric analysis is In the seminar we propose and use a unique tool,
complex because it entails several steps that employ developed in the R language, which follows a classic
numerous and diverse analyses and mapping software logical bibliometric workflow that we reconstruct.
tools, which are frequently available only under
commercial licenses. We have designed and produced an R-tool for
comprehensive bibliometric analyses. R is a language
These difficulties are compounded by the reality that and environment for statistical computing and
few researchers and practitioners are trained in how to graphics. It provides a wide variety of statistical and
review literature and to identify evidence-based graphical techniques and is highly extensible. In
practices. addition to enabling statistical operations, it is an
object-oriented and functional programming
The cumbersome nature of the process reduces the language; hence, you can automate your analyses and
possibilities and the potential of bibliometrics, create new functions. It has an open-software nature,
especially for scholars who have no general which means it is well supported by the user
programming skills. community and new functions are regularly
contributed by users, many of whom are prominent
Recently, automated workflows to assemble specialized statisticians.
software into a comprehensive and organized data flow
have begun to emerge for bibliometrics. They are As it is programmed in R, the proposed tool is flexible,
particularly well suited to multi-step analyses using can be rapidly upgraded, and can be integrated with
different types of software tools. other statistical R-packages. It is therefore useful in a
constantly changing field such as bibliometrics.
Visits 13.481 (last 12 months at March, 2017)

Aria, Cuccurullo (2017), JoI

Bibliometrix in Chinese
Recommended workflow
for science mapping
Study
design

• Data retrieval (Database)

Data • Data loading and converting
Interpretation
collection • Data cleaning.

• Software tools for science mapping

• R-packages for bibliometric analysis

Data • Network extraction

Data
• Data normalization
visualization Analysis • Data reduction
 Scientific document is the basic unit of a
complex relational system Co-citations

Collaborations
Word
co-occurrences
Bibliometrix
Study Design
Data collection: Main steps
Data retrieval Data downloading

Bibliographic dataframe Data importing and converting

Doc Authors Title Abstract Source Keywords Affilaition …
Bibliographic dataframe: an example Field tags
 Data collection
PRISMA diagram

• Keywords for query (Boolean operators)

• Timespan & timeslices
• Language (English)
• Types of documents (articles, …)
• Subject Categories (Mgmt, Fin, Ops, …)
• Sources (ABS, 2015; one-journal or …)
 Data Analysis
Intellectual
structure

Conceptual
Structure
(research front)

Coupling
Two works (A & B) refer
to a common work (a)

Co-citation
Two works (a & b) are cited
together by a common work (A)
Main functions
Software assisted bibliometrix
Description
workflow steps functions
• readFiles() • Loads a sequence of Scopus and Clarivate Analytics WoS export files into R
• Convert2df() • Creates a bibliographic data frame
Data loading and converting
• Uses Scopus API search to obtain information regarding documents on a set of
• retrievalByAuthorID()
authors using Scopus ID
• biblioAnalysis() • Returns an object of class bibliometrix
• summary() and plot() • Summarize the main results of the bibliometric analysis
• citations() • Identifies the most cited references or authors
• localCitations() • Identifies the most cited local authors
Descriptive bibliometric
• dominance() • Calculates the authors’ dominance ranking
analysis
• Hindex() • Measures productivity and citation impact of a scholar
• lotka() • Estimates Lotka’s law coefficients for scientific productivity
• keywordGrowth() • Calculates yearly cumulative occurrences of top keywords/terms
• keywordAssociation() • Associates authors' keywords to keywords plus
• metaTagExtraction() • Extracts other field tags, different from the standard WoS/Scopus codify
Document x Attribute matrix • Extracts and stems terms from textual fields (abstract, title, author's keywords, and
• termExtraction()
creation others) of a bibliographic data frame
• cocMatrix() • Computes a Document x Attribute matrix
• Calculates association strength, inclusion index, Jaccard’s coefficient, and Salton’s
Normalization • normalizeSimilarity()
similarity coefficient among objects of a bibliographic network
Data Reduction • conceptualStructure() • Creates conceptual structure map of a scientific field using MCA and Clustering
• Calculates the most frequently used bibliographic coupling, co-citation,
• biblioNetwork()
Network matrix creation collaboration, and co-occurrence networks
• histNetwork() • Creates a historical co-citation network from a bibliographic data frame
• networkPlot() • Plots a bibliographic network using internal R library or VOSviewer software
Mapping • histPlot() • Plots a historical co-citation network
• conceptualStructure() • Plots conceptual structure map of a scientific field using MCA and Clustering
Descriptive analysis
Wip on Big Data
 Matrix “Document x Attribute”
• Document’s attributes are connected to each other through the Doc itself: author(s) to journal,
keywords to publication date, etc.
• An attribute is an item of information associated to the document and stored in a field tag within
the bibliometric data frame (e.g., authors, publication source, keywords, cited references,
affiliations).
• These connections of different attributes generate a binary rectangular matrices (Document x
Attribute) that, in some cases, it can be represented as a bipartite networks
• Furthermore, scientific publications regularly contain references to other scientific works. This
generates a further network, namely, co-citation or coupling network
• These networks are analyzed in order to capture meaningful properties of the underlying research
system, and in particular to determine the influence of bibliometric units such as scholars and
journals.
Matrix 𝐷𝑜𝑐𝑢𝑚𝑒𝑛𝑡 × 𝑅𝑒𝑓𝑒𝑟𝑒𝑛𝑐𝑒
(cocMatrix function)
matrix 𝑨 𝐷𝑜𝑐𝑢𝑚𝑒𝑛𝑡 × 𝑅𝑒𝑓𝑒𝑟𝑒𝑛𝑐𝑒
Cited documents
A

Ref X ref Y Ref Z B

Doc A 1 0 1 X
Doc B 0 1 0 C
Doc C 1 0 1 Y
D
Doc D 0 1 0
E Z
Doc E 0 0 1
Doc F 1 1 0 F

Doc G 0 0 1 G

Citing documents Bipartite graph

Bibliographic Networks
(biblioNetwork function)

• Bibliographic coupling 𝐵𝑐𝑜𝑢𝑝 = 𝐴𝐶𝑅 × 𝐴′𝐶𝑅

• Co-citation 𝐵𝑐𝑜𝑐𝑖𝑡 = 𝐴′𝐶𝑅 × 𝐴𝐶𝑅

′
• Collaboration 𝐵𝑐𝑜𝑙𝑙 = 𝐴𝐴𝑈 × 𝐴𝐴𝑈
(among: authors, univ, dept, countries)

• Co-word 𝐵𝑐𝑜𝑐 = 𝐴′𝐼𝐷 × 𝐴𝐼𝐷

Co-citation coupling
“Co-citation Coupling” is the mirror image of “Bibliographic coupling”

• Co-citation coupling is a method used to establish a subject similarity

between two documents.
• If papers A and B are both cited by paper C, they may be said to be related
to one another, even though they don't directly cite each other.
• If papers A and B are both cited by many other papers, they have a stronger
relationship. The more papers they are cited by, the stronger their
relationship is.
Co-citation network
• A coupling network can be obtained using the general formulation:

𝐵𝑐𝑜𝑐𝑖𝑡 = 𝐴′𝐶𝑅 × 𝐴𝐶𝑅

• Like matrix 𝐵𝑐𝑜𝑢𝑝 , matrix 𝐵𝑐𝑜𝑐𝑖𝑡 is also symmetric.
• The main diagonal of 𝐵𝑐𝑜𝑐𝑖𝑡 contains the number of cases in which a
reference is cited in our dataframe.
• In other words, the diagonal element 𝐵𝑖𝑖 is the number of local
citations of the reference 𝑖.
 Co-citation analysis
′
𝑨𝑪𝑹 𝑨𝑪𝑹
Ref X Ref Y Ref Z
Doc A Doc B Doc C Doc D Doc E Doc F Doc G Doc A 1 0 1
Doc B 0 1 0
Ref X 1 0 1 0 0 1 0
× Doc C 1 0 1
Ref Y 0 1 0 1 0 1 0
Doc D 0 1 0
Ref Z 1 0 1 0 1 0 1 Doc E 0 0 1
Doc F 1 1 0
Doc G 0 0 1
Co-citation analysis (2)

Matrix 𝑩𝒄𝒐𝒄𝒊𝒕
X Y
X Y Z
X 3 1 2
Y 1 3 0
Z 2 0 4
Z
3 1 2 Degree

Co-citation Network
Document co-citation

• Position, proximity and bubble diameter

• Clusters
• Strenght of linkages
• Bridge papers
Journal co-citation
Collaboration network
 Historiograph

• Historiographic analysis generates

chronological tables as well as historiographs
which highlight the most-cited works in and
outside the collection.

• It will be used to help scholars quickly identify

the most significant work on a topic and trace its
year-by-year historical development.
Co-word analysis
through networks
Co-word Analysis
through MCA
 MOTOR
Thematic map THEMES

THEMATIC
NETWORK
HIGHLY DEVELOPED AND ISOLATED
THEMES
(NICHES)

BASIC AND TRANSVERSAL

EMERGING OR DECLINING THEMES
THEMES
What’s next?
• Shiny
• Lab of bibliometrics and data-knowledge discovery
• Bibliometrix R community
• Bibliometrix social (Follow us!)
• https://www.facebook.com/bibliometrix/
• https://twitter.com/search?q=%23bibliometrix&src=typd
• We are already working on new developments. They concern
• the extension of compatibility with other bibliographic databases such as PubMed
• The search of grey literature
• the improvement of reference disambiguation by string metric-based algorithms
• the introduction of direct citation and tri-citation analysis
• the use of hybrid methods that combine bibliometric and semantic approaches. The last-mentioned
development includes term-burst detection through expectile smoothing, thematic mapping and
evolution and latent semantic analysis