A brief introduction to bibliometrix
Massimo Aria and Corrado Cuccurullo
bibliometrix
Latest version
## bibliometrix 4.0.0
Citation for package ‘bibliometrix’
To cite bibliometrix in publications, please use:
Aria, M. & Cuccurullo, C. (2017) bibliometrix: An R-tool for comprehensive science mapping analysis, Journal of Informetrics, 11(4), pp 959-975, Elsevier.
A BibTeX entry for LaTeX users is
@Article{,
author = {Massimo Aria and Corrado Cuccurullo},
title = {bibliometrix: An R-tool for comprehensive science mapping analysis},
journal = {Journal of Informetrics},
volume = {11},
number = {4},
pages = {959-975},
publisher = {Elsevier},
year = {2017},
url = {https://doi.org/10.1016/j.joi.2017.08.007},
}
Introduction
bibliometrix package provides a set of tools for quantitative research in bibliometrics and scientometrics.
Bibliometrics turns the main tool of science, quantitative analysis, on itself. Essentially, bibliometrics is the application of quantitative analysis and statistics to publications such as journal articles and their accompanying citation counts. Quantitative evaluation of publication and citation data is now used in almost all scientific fields to evaluate growth, maturity, leading authors, conceptual and intellectual maps, trends of a scientific community.
Bibliometrics is also used in research performance evaluation, especially in university and government labs, and also by policymakers, research directors and administrators, information specialists and librarians, and scholars themselves.
bibliometrix supports scholars in three key phases of analysis:
Data importing and conversion to R format;
Bibliometric analysis of a publication dataset;
Building matrices for co-citation, coupling, collaboration, and co-word analysis. Matrices are the input data for performing network analysis, multiple correspondence analysis, and any other data reduction techniques.
Bibliographic databases
bibliometrix works with data extracted from the four main bibliographic databases: SCOPUS, Clarivate Analytics Web of Science, Cochrane Database of Systematic Reviews (CDSR) and RISmed PubMed/MedLine.
SCOPUS (https://www.scopus.com), founded in 2004, offers a great deal of flexibility for the bibliometric user. It permits to query for different fields, such as titles, abstracts, keywords, references and so on. SCOPUS allows for relatively easy downloading data-queries, although there are some limits on very large results sets with over 2,000 items.
Clarivate Analytics Web of Science (WoS) (https://www.webofknowledge.com), owned by Clarivate
Analytics, was founded by Eugene Garfield, one of the pioneers of
bibliometrics.
This platform includes many different collections.
Cochrane Database of Systematic Reviews (https://www.cochranelibrary.com/cdsr/about-cdsr) is the leading resource for systematic reviews in health care. The CDSR includes Cochrane Reviews (the systematic reviews) and protocols for Cochrane Reviews as well as editorials. The CDSR also has occasional supplements. The CDSR is updated regularly as Cochrane Reviews are published “when ready” and form monthly issues; see publication schedule.
PubMed comprises more than 28 million citations for biomedical literature from MEDLINE, life science journals, and online books. Citations may include links to full-text content from PubMed Central and publisher websites.
Data acquisition
Bibliographic data may be obtained by querying the SCOPUS or Clarivate Analytics Web of Science (WoS) database by diverse fields, such as topic, author, journal, timespan, and so on.
In this example, we show how to download data, querying a term in the manuscript title field.
We choose the generic term “bibliometrics”.
Querying from Clarivate Analytics WoS
At the link https://www.webofknowledge.com, select Web of Science Core Collection database.
Write the keyword “bibliometrics” in the search field and select title from the drop-down menu (see figure 1).
Figure 1
Choose SCI-EXPANDED and SSCI citation indexes.
The search yielded 291 results on May 09, 2016.
Results can be refined using options on the left side of the page (the type of manuscript, source, subject category, etc.).
After refining the query, you can add records to your Marked List by clicking the button “add to marked list” at the end of the page and selecting the records to save (see figure 2).
Figure 2
The Marked List page provides you with a list of publications selected and various means of exporting data.
To export the data you desire, choose the export tool and follow the three intuitive steps (see figure 3).
Figure 3
The export tool allows you to select the diverse fields to save. So, select the fields you are interested in (for example all the available data about marked records).
To download an export file, in an appropriate format for the bibliometrix package, make sure to select the option “Save to Other File Formats” and choose Bibtex or Plain Text.
The WoS platform permits to export only 500 records at a time.
The Clarivate Analytics Web of Science export tool creates an export file with a default name “savedrecs” with an extension “.txt” or “.bib” for plain text or BibTeX format respectively. Export files can be separately stored.
Querying from SCOPUS
The access to SCOPUS is via https://www.scopus.com.
To find all articles whose title includes the term “bibliometrics”, simply write this keyword in the field and select “Article Title” (see figure 4)
Figure 4
The search yielded 414 results on May 09, 2016.
You can download the references (up to 2,000 full records) by checking the ‘Select All’ box and clicking on the link ‘Export’. Choose the file type “BibTeX export” and “all available information” (see figure 5).
Figure 5
The SCOPUS export tool creates an export file with the default name “scopus.bib”.
bibliometrix installation
Download and install the most recent version of R (https://cran.r-project.org)
Download and install the most recent version of Rstudio (https://rstudio.com)
Open Rstudio, in the console window, digit:
install.packages(“bibliometrix”, dependencies=TRUE) ### installs bibliometrix package and dependencies
library(bibliometrix) ### load bibliometrix package## To cite bibliometrix in publications, please use:
##
## Aria, M. & Cuccurullo, C. (2017) bibliometrix: An R-tool for comprehensive science mapping analysis,
## Journal of Informetrics, 11(4), pp 959-975, Elsevier.
##
##
## https://www.bibliometrix.org
##
##
## For information and bug reports:
## - Send an email to info@bibliometrix.org
## - Write a post on https://github.com/massimoaria/bibliometrix/issues
##
## Help us to keep Bibliometrix free to download and use by contributing with a small donation to support our research team (https://bibliometrix.org/donate.html)
##
##
## To start with the shiny web-interface, please digit:
## biblioshiny()Data loading and converting
The export file can be read and converted using by R using the function convert2df:
convert2df(file, dbsource, format)
The argument file is a character vector containing the name of export files downloaded from SCOPUS, Clarivate Analytics WOS, Digital Science Dimensions, PubMed or Cochrane CDSR website. file can also contains the name of a json/xlm object download using Digital Science Dimenions or PubMed APIs (through the packages dimensionsR and pubmedR.
es. file <- c(“file1.txt”,“file2.txt”, …)
file <- "https://www.bibliometrix.org/datasets/savedrecs.bib"
M <- convert2df(file = file, dbsource = "isi", format = "bibtex")##
## Converting your isi collection into a bibliographic dataframe
##
## Done!
##
##
## Generating affiliation field tag AU_UN from C1: Done!convert2df creates a bibliographic data frame with cases corresponding to manuscripts and variables to Field Tag in the original export file.
convert2df accepts two additional arguments: dbsource and format.
The argument dbsource indicates from which database the collection has been downloaded.
It can be:
“isi” or “wos” (for Clarivate Analytics Web of Science database),
“scopus” (for SCOPUS database),
“dimensions” (for DS Dimensions database)
“pubmed” (for PubMed/Medline database),
“cochrane” (for Cochrane Library database of systematic reviews).
The argument format indicates the file format of the imported collection. It can be “plaintext” or “bibtex” for WOS collection and mandatorily “bibtext” for SCOPUS collection. The argument is ignored if the collection comes from Pubmed or Cochrane.
Each manuscript contains several elements, such as authors’ names, title, keywords and other information. All these elements constitute the bibliographic attributes of a document, also called metadata.
Data frame columns are named using the standard Clarivate Analytics WoS Field Tag codify.
The main field tags are:
| Field Tag | Description |
|---|---|
| AU | Authors |
| TI | Document Title |
| SO | Publication Name (or Source) |
| JI | ISO Source Abbreviation |
| DT | Document Type |
| DE | Authors’ Keywords |
| ID | Keywords associated by SCOPUS or ISI database |
| AB | Abstract |
| C1 | Author Address |
| RP | Reprint Address |
| CR | Cited References |
| TC | Times Cited |
| PY | Year |
| SC | Subject Category |
| UT | Unique Article Identifier |
| DB | Bibliographic Database |
For a complete list of field tags see https://www.bibliometrix.org/documents/Field_Tags_bibliometrix.pdf
Bibliometric Analysis
The first step is to perform a descriptive analysis of the bibliographic data frame.
The function biblioAnalysis calculates main bibliometric measures using this syntax:
results <- biblioAnalysis(M, sep = ";")The function biblioAnalysis returns an object of class “bibliometrix”.
An object of class “bibliometrix” is a list containing the following components:
| List element | Description |
|---|---|
| Articles | the total number of manuscripts |
| Authors | the authors’ frequency distribution |
| AuthorsFrac | the authors’ frequency distribution (fractionalized) |
| FirstAuthors | corresponding author of each manuscript |
| nAUperPaper | the number of authors per manuscript |
| Appearances | the number of author appearances |
| nAuthors | the number of authors |
| AuMultiAuthoredArt | the number of authors of multi-authored articles |
| MostCitedPapers | the list of manuscripts sorted by citations |
| Years | publication year of each manuscript |
| FirstAffiliation | the affiliation of the corresponding author |
| Affiliations | the frequency distribution of affiliations (of all co-authors for each paper) |
| Aff_frac | the fractionalized frequency distribution of affiliations (of all co-authors for each paper) |
| CO | the affiliation country of the corresponding author |
| Countries | the affiliation countries’ frequency distribution |
| CountryCollaboration | the intra-country (SCP) and inter-country (MCP) collaboration indices |
| TotalCitation | the number of times each manuscript has been cited |
| TCperYear | the yearly average number of times each manuscript has been cited |
| Sources | the frequency distribution of sources (journals, books, etc.) |
| DE | the frequency distribution of authors’ keywords |
| ID | the frequency distribution of keywords associated to the manuscript by SCOPUS and Thomson Reuters’ ISI Web of Knowledge databases |
Functions summary and plot
To summarize main results of the bibliometric analysis, use the generic function summary. It displays main information about the bibliographic data frame and several tables, such as annual scientific production, top manuscripts per number of citations, most productive authors, most productive countries, total citation per country, most relevant sources (journals) and most relevant keywords.
Main information table describes the collection size in terms of number of documents, number of authors, number of sources, number of keywords, timespan, and average number of citations.
Furthermore, many different co-authorship indices are shown. In particular, the Authors per Article index is calculated as the ratio between the total number of authors and the total number of articles. The Co-Authors per Articles index is calculated as the average number of co-authors per article. In this case, the index takes into account the author appearances while for the “authors per article” an author, even if he has published more than one article, is counted only once. For that reasons, Authors per Article index \(\le\) Co-authors per Article index.
The Collaboration Index (CI) is calculated as Total Authors of Multi-Authored Articles/Total Multi-Authored Articles (Elango and Rajendran, 2012; Koseoglu, 2016). In other word, the Collaboration Index is a Co-authors per Article index calculated only using the multi-authored article set.
Elango, B., & Rajendran, P. (2012). Authorship trends and collaboration pattern in the marine sciences literature: a scientometric study. International Journal of Information Dissemination and Technology, 2(3), 166.
Koseoglu, M. A. (2016). Mapping the institutional collaboration network of strategic management research: 1980–2014. Scientometrics, 109(1), 203-226.
summary accepts two additional arguments. k is a formatting value that indicates the number of rows of each table. pause is a logical value (TRUE or FALSE) used to allow (or not) pause in screen scrolling. Choosing k=10 you decide to see the first 10 Authors, the first 10 sources, etc.
options(width=100)
S <- summary(object = results, k = 10, pause = FALSE)##
##
## MAIN INFORMATION ABOUT DATA
##
## Timespan 1985 : 2015
## Sources (Journals, Books, etc) 141
## Documents 291
## Annual Growth Rate % 7.18
## Document Average Age 16.7
## Average citations per doc 11.73
## Average citations per year per doc 0.6489
## References 6768
##
## DOCUMENT TYPES
## art exhibit review 1
## article 160
## article; proceedings paper 7
## biographical-item 1
## book review 32
## correction, addition 1
## editorial material 41
## letter 16
## meeting abstract 4
## note 3
## review 25
##
## DOCUMENT CONTENTS
## Keywords Plus (ID) 475
## Author's Keywords (DE) 365
##
## AUTHORS
## Authors 523
## Author Appearances 635
## Authors of single-authored docs 121
##
## AUTHORS COLLABORATION
## Single-authored docs 144
## Documents per Author 0.556
## Co-Authors per Doc 2.18
## International co-authorships % 12.37
##
##
## Annual Scientific Production
##
## Year Articles
## 1985 4
## 1986 3
## 1987 6
## 1988 7
## 1989 8
## 1990 6
## 1991 7
## 1992 6
## 1993 5
## 1994 7
## 1995 1
## 1996 8
## 1997 4
## 1998 5
## 1999 2
## 2000 7
## 2001 8
## 2002 5
## 2003 1
## 2004 3
## 2005 12
## 2006 5
## 2007 5
## 2008 8
## 2009 14
## 2010 17
## 2011 20
## 2012 25
## 2013 21
## 2014 29
## 2015 32
##
## Annual Percentage Growth Rate 7.177346
##
##
## Most Productive Authors
##
## Authors Articles Authors Articles Fractionalized
## 1 BORNMANN L 8 BORNMANN L 4.67
## 2 KOSTOFF RN 8 WHITE HD 3.50
## 3 MARX W 6 MARX W 3.17
## 4 HUMENIK JA 5 ATKINSON R 3.00
## 5 ABRAMO G 4 BROADUS RN 3.00
## 6 D'ANGELO CA 4 CRONIN B 3.00
## 7 GARG KC 4 BORGMAN CL 2.50
## 8 GLANZEL W 4 MCCAIN KW 2.50
## 9 WHITE HD 4 PERITZ BC 2.50
## 10 ATKINSON R 3 KOSTOFF RN 2.10
##
##
## Top manuscripts per citations
##
## Paper DOI TC TCperYear NTC
## 1 DAIM TU, 2006, TECHNOL FORECAST SOC CHANG 10.1016/j.techfore.2006.04.004 211 12.41 3.44
## 2 WHITE HD, 1989, ANNU REV INFORM SCI TECHNOL NA 196 5.76 5.58
## 3 BORGMAN CL, 2002, ANNU REV INFORM SCI TECHNOL NA 192 9.14 3.60
## 4 WEINGART P, 2005, SCIENTOMETRICS 10.1007/s11192-005-0007-7 151 8.39 6.79
## 5 NARIN F, 1994, SCIENTOMETRICS 10.1007/BF02017219 141 4.86 4.72
## 6 CRONIN B, 2001, J INF SCI NA 129 5.86 2.78
## 7 CHEN YC, 2011, SCIENTOMETRICS 10.1007/s11192-010-0289-2 101 8.42 5.92
## 8 HOOD WW, 2001, SCIENTOMETRICS 10.1023/A:1017919924342 71 3.23 1.53
## 9 D'ANGELO CA, 2011, J AM SOC INF SCI TECHNOL 10.1002/asi.21460 64 5.33 3.75
## 10 NARIN F, 1994, EVAL REV 10.1177/0193841X9401800107 62 2.14 2.08
##
##
## Corresponding Author's Countries
##
## Country Articles Freq SCP MCP MCP_Ratio
## 1 USA 81 0.3057 76 5 0.0617
## 2 UNITED KINGDOM 27 0.1019 27 0 0.0000
## 3 GERMANY 17 0.0642 12 5 0.2941
## 4 FRANCE 13 0.0491 11 2 0.1538
## 5 BRAZIL 12 0.0453 10 2 0.1667
## 6 CHINA 10 0.0377 8 2 0.2000
## 7 INDIA 10 0.0377 10 0 0.0000
## 8 AUSTRALIA 8 0.0302 6 2 0.2500
## 9 CANADA 8 0.0302 7 1 0.1250
## 10 SPAIN 8 0.0302 8 0 0.0000
##
##
## SCP: Single Country Publications
##
## MCP: Multiple Country Publications
##
##
## Total Citations per Country
##
## Country Total Citations Average Article Citations
## 1 USA 1831 22.60
## 2 GERMANY 330 19.41
## 3 ITALY 163 32.60
## 4 AUSTRALIA 134 16.75
## 5 UNITED KINGDOM 125 4.63
## 6 CANADA 111 13.88
## 7 INDIA 85 8.50
## 8 IRAN 74 37.00
## 9 SPAIN 73 9.12
## 10 BELGIUM 70 10.00
##
##
## Most Relevant Sources
##
## Sources Articles
## 1 SCIENTOMETRICS 49
## 2 JOURNAL OF THE AMERICAN SOCIETY FOR INFORMATION SCIENCE AND TECHNOLOGY 14
## 3 JOURNAL OF THE AMERICAN SOCIETY FOR INFORMATION SCIENCE 8
## 4 JOURNAL OF DOCUMENTATION 6
## 5 JOURNAL OF INFORMATION SCIENCE 6
## 6 JOURNAL OF INFORMETRICS 6
## 7 BRITISH JOURNAL OF ANAESTHESIA 5
## 8 LIBRI 5
## 9 SOCIAL WORK IN HEALTH CARE 5
## 10 TECHNOLOGICAL FORECASTING AND SOCIAL CHANGE 5
##
##
## Most Relevant Keywords
##
## Author Keywords (DE) Articles Keywords-Plus (ID) Articles
## 1 BIBLIOMETRICS 63 SCIENCE 38
## 2 CITATION ANALYSIS 11 INDICATORS 24
## 3 SCIENTOMETRICS 7 IMPACT 23
## 4 IMPACT FACTOR 5 CITATION 20
## 5 INFORMATION RETRIEVAL 5 CITATION ANALYSIS 15
## 6 PEER REVIEW 5 JOURNALS 14
## 7 CITATION 4 H-INDEX 13
## 8 CITATIONS 4 PUBLICATION 12
## 9 H-INDEX 4 INFORMATION-SCIENCE 10
## 10 IMPACT FACTORS 4 IMPACT FACTORS 8Some basic plots can be drawn using the generic function :
plot(x = results, k = 10, pause = FALSE)
Analysis of Cited References
The function citations generates the frequency table of the most cited references or the most cited first authors (of references).
For each manuscript, cited references are in a single string stored in the column “CR” of the data frame.
For a correct extraction, you need to identify the separator field
among different references, used by ISI or SCOPUS database. Usually, the
default separator is “;” or ". " (a dot with double
space).
# M$CR[1]The figure shows the reference string of the first manuscript. In
this case, the separator field is sep = ";".
Figure 6
To obtain the most frequent cited manuscripts:
CR <- citations(M, field = "article", sep = ";")
cbind(CR$Cited[1:10])## [,1]
## HIRSCH JE, 2005, P NATL ACAD SCI USA, V102, P16569, DOI 10.1073/PNAS.0507655102 29
## PRITCHAR.A, 1969, J DOC, V25, P348 22
## SMALL H, 1973, J AM SOC INFORM SCI, V24, P265, DOI 10.1002/ASI.4630240406 20
## DE SOLLA PRICE DJ, 1963, LITTLE SCI BIG SCI 19
## BRADFORD S. C, 1934, ENGINEERING-LONDON, V137, P85 14
## GARFIELD E, 2006, JAMA-J AM MED ASSOC, V295, P90, DOI 10.1001/JAMA.295.1.90 12
## MOED H. F., 2005, CITATION ANAL RES EV 12
## COLE FRANCIS J., 1917, SCI PROGR, V11, P578 11
## GROSS P L, 1927, SCIENCE, V66, P385, DOI 10.1126/SCIENCE.66.1713.385 11
## LOTKA A. Y., 1926, J WASHINGTON ACAD SC, V16, P317 11To obtain the most frequent cited first authors:
CR <- citations(M, field = "author", sep = ";")
cbind(CR$Cited[1:10])## [,1]
## GARFIELD E 167
## KOSTOFF RN 129
## BORNMANN L 92
## SMALL H 75
## WHITE HD 74
## CRONIN B 67
## NARIN F 51
## GLANZEL W 50
## LEYDESDORFF L 47
## EGGHE L 45The function localCitations generates the frequency table of the most local cited authors. Local citations measure how many times an author (or a document) included in this collection have been cited by other authors also in the collection.
To obtain the most frequent local cited authors:
CR <- localCitations(M, sep = ";")
CR$Authors[1:10,]## Author LocalCitations
## 21 BARKER K 9
## 190 HOLDEN G 9
## 386 ROSENBERG G 9
## 1 ABRAMO G 8
## 83 D'ANGELO CA 8
## 50 BROADUS RN 7
## 199 HUMENIK JA 6
## 233 KOSTOFF RN 6
## 354 PFEIL KM 6
## 455 TSHITEYA R 6CR$Papers[1:10,]## Paper DOI Year LCS GCS
## 9 BROADUS RN, 1987, SCIENTOMETRICS 10.1007/BF02016680 1987 5 38
## 108 HOLDEN G, 2005, SOC WORK HEALTH CARE 10.1300/J010v41n03\\_01 2005 5 22
## 141 ABRAMO G, 2009, RES POLICY 10.1016/j.respol.2008.11.001 2009 5 43
## 45 SENGUPTA IN, 1992, LIBRI 10.1515/libr.1992.42.2.75 1992 4 20
## 83 CRONIN B, 2000, J DOC 10.1108/EUM0000000007123 2000 4 20
## 95 KOSTOFF RN, 2002, J POWER SOURCES 10.1016/S0378-7753(02)00233-1 2002 4 34
## 109 HOLDEN G, 2005, SOC WORK HEALTH CARE-a 10.1300/J010v41n03\\_03 2005 4 34
## 6 PERSSON O, 1986, SCIENTOMETRICS 10.1007/BF02016861 1986 3 15
## 7 DEGLAS F, 1986, LIBRI 10.1515/libr.1986.36.1.40 1986 2 8
## 10 BROADUS RN, 1987, J EDUC LIBR INF SCI 10.2307/40323625 1987 2 2Top-Authors’ Productivity over the Time
The function AuthorProdOverTime calculates and plots the authors’ production (in terms of number of publications, and total citations per year) over the time.
Function arguments are: M a bibliographic data frame; k is the number of k Top Authors; graph is a logical. If graph=TRUE, the function plots the author production over time graph.
topAU <- authorProdOverTime(M, k = 10, graph = TRUE)
## Table: Author's productivity per year
head(topAU$dfAU)## Author year freq TC TCpY
## 1 ABRAMO G 2009 1 43 3.0714286
## 2 ABRAMO G 2011 3 115 9.5833333
## 3 ATKINSON R 2012 3 0 0.0000000
## 4 BORNMANN L 2012 1 2 0.1818182
## 5 BORNMANN L 2013 3 39 3.9000000
## 6 BORNMANN L 2014 2 4 0.4444444## Table: Auhtor's documents list
#head(topAU$dfPapersAU)Lotka’s Law coefficient estimation
The function lotka estimates Lotka’s law coefficients for scientific productivity (Lotka A.J., 1926).
Lotka’s law describes the frequency of publication by authors in any given field as an inverse square law, where the number of authors publishing a certain number of articles is a fixed ratio to the number of authors publishing a single article. This assumption implies that the theoretical beta coefficient of Lotka’s law is equal to 2.
Using lotka function is possible to estimate the Beta coefficient of our bibliographic collection and assess, through a statistical test, the similarity of this empirical distribution with the theoretical one.
L <- lotka(results)
# Author Productivity. Empirical Distribution
L$AuthorProd## N.Articles N.Authors Freq
## 1 1 453 0.866156788
## 2 2 48 0.091778203
## 3 3 13 0.024856597
## 4 4 5 0.009560229
## 5 5 1 0.001912046
## 6 6 1 0.001912046
## 7 8 2 0.003824092# Beta coefficient estimate
L$Beta## [1] 3.043039# Constant
L$C## [1] 0.688399# Goodness of fit
L$R2## [1] 0.917464# P-value of K-S two sample test
L$p.value## [1] 0.2031888The table L$AuthorProd shows the observed distribution of scientific productivity in our example.
The estimated Beta coefficient is 3.05 with a goodness of fit equal to 0.94. Kolmogorov-Smirnoff two sample test provides a p-value 0.09 that means there is not a significant difference between the observed and the theoretical Lotka distributions.
You can compare the two distributions using plot function:
# Observed distribution
Observed=L$AuthorProd[,3]
# Theoretical distribution with Beta = 2
Theoretical=10^(log10(L$C)-2*log10(L$AuthorProd[,1]))
plot(L$AuthorProd[,1],Theoretical,type="l",col="red",ylim=c(0, 1), xlab="Articles",ylab="Freq. of Authors",main="Scientific Productivity")
lines(L$AuthorProd[,1],Observed,col="blue")
legend(x="topright",c("Theoretical (B=2)","Observed"),col=c("red","blue"),lty = c(1,1,1),cex=0.6,bty="n")
Bibliographic network matrices
Manuscript’s attributes are connected to each other through the manuscript itself: author(s) to journal, keywords to publication date, etc.
These connections of different attributes generate bipartite networks that can be represented as rectangular matrices (Manuscripts x Attributes).
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.
Bipartite networks
cocMatrix is a general function to compute a bipartite network selecting one of the metadata attributes.
For example, to create a network Manuscript x Publication Source you have to use the field tag “SO”:
A <- cocMatrix(M, Field = "SO", sep = ";")A is a rectangular binary matrix, representing a bipartite network where rows and columns are manuscripts and sources respectively.
The generic element \(a_{ij}\) is 1 if the manuscript \(i\) has been published in source \(j\), 0 otherwise.
The \(j-th\) column sum \(a_j\) is the number of manuscripts published in source \(j\).
Sorting, in decreasing order, the column sums of A, you can see the most relevant publication sources:
sort(Matrix::colSums(A), decreasing = TRUE)[1:5]## SCIENTOMETRICS
## 49
## JOURNAL OF THE AMERICAN SOCIETY FOR INFORMATION SCIENCE AND TECHNOLOGY
## 14
## JOURNAL OF THE AMERICAN SOCIETY FOR INFORMATION SCIENCE
## 8
## JOURNAL OF DOCUMENTATION
## 6
## JOURNAL OF INFORMATION SCIENCE
## 6Following this approach, you can compute several bipartite networks:
- Citation network
A <- cocMatrix(M, Field = "CR", sep = ". ")- Author network
A <- cocMatrix(M, Field = "AU", sep = ";")- Country network
Authors’ Countries is not a standard attribute of the bibliographic data frame. You need to extract this information from affiliation attribute using the function metaTagExtraction.
M <- metaTagExtraction(M, Field = "AU_CO", sep = ";")
# A <- cocMatrix(M, Field = "AU_CO", sep = ";")metaTagExtraction allows to extract the following additional
field tags: Authors’ countries (Field = "AU_CO");
First Author’s countries (Field = "AU_CO");
First author of each cited reference
(Field = "CR_AU"); Publication source of each cited
reference (Field = "CR_SO"); and Authors’
affiliations (Field = "AU_UN").
- Author keyword network
A <- cocMatrix(M, Field = "DE", sep = ";")- Keyword Plus network
A <- cocMatrix(M, Field = "ID", sep = ";")- Etc.
Bibliographic coupling
Two articles are said to be bibliographically coupled if at least one cited source appears in the bibliographies or reference lists of both articles (Kessler, 1963).
A coupling network can be obtained using the general formulation:
\[ B = A \times A^T \] where A is a bipartite network.
Element \(b_{ij}\) indicates how many bibliographic couplings exist between manuscripts \(i\) and \(j\). In other words, \(b_{ij}\) gives the number of paths of length 2, via which one moves from \(i\) along the arrow and then to \(j\) in the opposite direction.
\(B\) is a symmetrical matrix \(B = B^T\).
The strength of the coupling of two articles, \(i\) and \(j\) is defined simply by the number of references that the articles have in common, as given by the element \(b_{ij}\) of matrix \(B\).
The function biblioNetwork calculates, starting from a bibliographic data frame, the most frequently used coupling networks: Authors, Sources, and Countries.
biblioNetwork uses two arguments to define the network to compute:
analysis argument can be “co-citation”, “coupling”, “collaboration”, or “co-occurrences”.
network argument can be “authors”, “references”, “sources”, “countries”, “universities”, “keywords”, “author_keywords”, “titles” and “abstracts”.
The following code calculates a classical article coupling network:
NetMatrix <- biblioNetwork(M, analysis = "coupling", network = "references", sep = ". ")Articles with only a few references, therefore, would tend to be more weakly bibliographically coupled, if coupling strength is measured simply according to the number of references that articles contain in common.
This suggests that it might be more practical to switch to a relative measure of bibliographic coupling.
normalizeSimilarity function calculates Association strength, Inclusion, Jaccard or Salton similarity among vertices of a network. normalizeSimilarity can be recalled directly from networkPlot() function using the argument normalize.
NetMatrix <- biblioNetwork(M, analysis = "coupling", network = "authors", sep = ";")
#net=networkPlot(NetMatrix, normalize = "salton", weighted=NULL, n = 100, Title = "Authors' Coupling", type = "fruchterman", size=5,size.cex=T,remove.multiple=TRUE,labelsize=0.8,label.n=10,label.cex=F)Bibliographic co-citation
We talk about co-citation of two articles when both are cited in a third article. Thus, co-citation can be seen as the counterpart of bibliographic coupling.
A co-citation network can be obtained using the general formulation:
\[ C = A^T \times A \] where A is a bipartite network.
Like matrix \(B\), matrix \(C\) is also symmetric. The main diagonal of \(C\) contains the number of cases in which a reference is cited in our data frame.
In other words, the diagonal element \(c_{i}\) is the number of local citations of the reference \(i\).
Using the function biblioNetwork, you can calculate a classical reference co-citation network:
# NetMatrix <- biblioNetwork(M, analysis = "co-citation", network = "references", sep = ". ")Bibliographic collaboration
Scientific collaboration network is a network where nodes are authors and links are co-authorships as the latter is one of the most well-documented forms of scientific collaboration (Glanzel, 2004).
An author collaboration network can be obtained using the general formulation:
\[ AC = A^T \times A \] where A is a bipartite network Manuscripts x Authors.
The diagonal element \(ac_{i}\) is the number of manuscripts authored or co-authored by researcher \(i\).
Using the function biblioNetwork, you can calculate an authors’ collaboration network:
# NetMatrix <- biblioNetwork(M, analysis = "collaboration", network = "authors", sep = ";")or a country collaboration network:
# NetMatrix <- biblioNetwork(M, analysis = "collaboration", network = "countries", sep = ";")Descriptive analysis of network graph characteristics
The function networkStat calculates several summary statistics.
In particular, starting from a bibliographic matrix (or an igraph object), two groups of descriptive measures are computed:
The summary statistics of the network;
The main indices of centrality and prestige of vertices.
# An example of a classical keyword co-occurrences network
NetMatrix <- biblioNetwork(M, analysis = "co-occurrences", network = "keywords", sep = ";")
netstat <- networkStat(NetMatrix)The summary statistics of the network
This group of statistics allows to describe the structural properties of a network:
Size is the number of vertices composing the network;
Density is the proportion of present edges from all possible edges in the network;
Transitivity is the ratio of triangles to connected triples;
Diameter is the longest geodesic distance (length of the shortest path between two nodes) in the network;
Degree distribution is the cumulative distribution of vertex degrees;
Degree centralization is the normalized degree of the overall network;
Closeness centralization is the normalized inverse of the vertex average geodesic distance to others in the network;
Eigenvector centralization is the first eigenvector of the graph matrix;
Betweenness centralization is the normalized number of geodesics that pass through the vertex;
Average path length is the mean of the shortest distance between each pair of vertices in the network.
names(netstat$network)## [1] "networkSize" "networkDensity" "networkTransitivity" "networkDiameter"
## [5] "networkDegreeDist" "networkCentrDegree" "networkCentrCloseness" "networkCentrEigen"
## [9] "networkCentrbetweenness" "NetworkAverPathLeng"The main indices of centrality and prestige of vertices
These measures help to identify the most important vertices in a network and the propensity of two vertices that are connected to be both connected to a third vertex.
The statistics, at vertex level, returned by networkStat are:
Degree centrality
Closeness centrality measures how many steps are required to access every other vertex from a given vertex;
Eigenvector centrality is a measure of being well-connected connected to the well-connected;
Betweenness centrality measures brokerage or gatekeeping potential. It is (approximately) the number of shortest paths between vertices that pass through a particular vertex;
PageRank score approximates probability that any message will arrive to a particular vertex. This algorithm was developed by Google founders, and originally applied to website links;
Hub Score estimates the value of the links outgoing from the vertex. It was initially applied to the web pages;
Authority Score is another measure of centrality initially applied to the Web. A vertex has high authority when it is linked by many other vertices that are linking many other vertices;
Vertex Ranking is an overall vertex ranking obtained as a linear weighted combination of the centrality and prestige vertex measures. The weights are proportional to the loadings of the first component of the Principal Component Analysis.
names(netstat$vertex)## NULLTo summarize the main results of the networkStat function, use the generic function summary. It displays the main information about the network and vertex description through several tables.
summary accepts one additional argument. k is a formatting value that indicates the number of rows of each table. Choosing k=10, you decide to see the first 10 vertices.
summary(netstat, k=10)##
##
## Main statistics about the network
##
## Size 475
## Density 0.024
## Transitivity 0.335
## Diameter 5
## Degree Centralization 0.301
## Average path length 2.743
## Visualizing bibliographic networks
All bibliographic networks can be graphically visualized or modeled.
Here, we show how to visualize networks using function networkPlot and VOSviewer software by Nees Jan van Eck and Ludo Waltman (https://www.vosviewer.com).
Using the function networkPlot, you can plot a network created by biblioNetwork using R routines or using VOSviewer.
The main argument of networkPlot is type. It indicates the network map layout: circle, kamada-kawai, mds, etc. Choosing type=“vosviewer”, the function automatically: (i) saves the network into a pajek network file, named “vosnetwork.net”; (ii) starts an instance of VOSviewer which will map the file “vosnetwork.net”. You need to declare, using argument vos.path, the full path of the folder where VOSviewer software is located (es. vos.path=‘c:/software/VOSviewer’).
Country Scientific Collaboration
# Create a country collaboration network
M <- metaTagExtraction(M, Field = "AU_CO", sep = ";")
NetMatrix <- biblioNetwork(M, analysis = "collaboration", network = "countries", sep = ";")
# Plot the network
net=networkPlot(NetMatrix, n = dim(NetMatrix)[1], Title = "Country Collaboration", type = "circle", size=TRUE, remove.multiple=FALSE,labelsize=0.7,cluster="none")
Co-Citation Network
# Create a co-citation network
# NetMatrix <- biblioNetwork(M, analysis = "co-citation", network = "references", sep = ";")
# Plot the network
#net=networkPlot(NetMatrix, n = 30, Title = "Co-Citation Network", type = "fruchterman", size=T, remove.multiple=FALSE, labelsize=0.7,edgesize = 5)Keyword co-occurrences
# Create keyword co-occurrences network
NetMatrix <- biblioNetwork(M, analysis = "co-occurrences", network = "keywords", sep = ";")
# Plot the network
net=networkPlot(NetMatrix, normalize="association", weighted=T, n = 30, Title = "Keyword Co-occurrences", type = "fruchterman", size=T,edgesize = 5,labelsize=0.7)
Co-Word Analysis: The conceptual structure of a field
The aim of the co-word analysis is to map the conceptual structure of a framework using the word co-occurrences in a bibliographic collection.
The analysis can be performed through dimensionality reduction techniques such as Multidimensional Scaling (MDS), Correspondence Analysis (CA) or Multiple Correspondence Analysis (MCA).
Here, we show an example using the function conceptualStructure that performs a CA or MCA to draw a conceptual structure of the field and K-means clustering to identify clusters of documents which express common concepts. Results are plotted on a two-dimensional map.
conceptualStructure includes natural language processing (NLP) routines (see the function termExtraction) to extract terms from titles and abstracts. In addition, it implements the Porter’s stemming algorithm to reduce inflected (or sometimes derived) words to their word stem, base or root form.
# Conceptual Structure using keywords (method="CA")
CS <- conceptualStructure(M,field="ID", method="CA", minDegree=4, clust=5, stemming=FALSE, labelsize=10, documents=10)
