Introduction

Evaluation and interpretation of scientific productions can be so helpful in determination of prominent authors, active departments and hot topics in a specific field. This process is called scientometrics and bibliometrix. Scientometrics refers to “all quantitative aspects science and scientific research” (Sengupta 1992). On the other hand, Bibliometrics refers to “the application of mathematics and statistical methods to books and other forms of written communication” (Pritchard 1969). Visualization and Statistical methods of these published documents can be analyzed using R bibliometrix package. This package is created and developed by [Massimo Aria] (https://masimoaria.com) and [Corrado Coccurullo] (https://www.corradococcurullo.com).

Our purpose is to investigate the RStudio applications of published scientific papers of PubMed database. In order to, we used bibliometrix package in Rstudio.

Call required packages

First of all we should install and call all of required packages for our analysis.

install.packages("bibliometrix")
install.packages("kableExtra")
install.packages("dplyr")
install.packages("ggplot2")
install.packages("reshape2")
install.packages("pubmedR")

options(scipen = 999)
library(bibliometrix)
library(kableExtra)
library(dplyr)
library(ggplot2)
library(reshape2)
library(pubmedR)

Search strategy

The search strategy should be considered based on a predefined search text

An API key is required to better and faster searches.Nevertheless, NULL can be put instead of a specific value.

As, RStudio and R programming language are interchangeably used, then both of them are considered in search strategy. After determining the search strategy we searched and finally found 649 documents based on searched terms including articles, book chapters, conference papers and etc.

Now, this data set is necessary to be converted to data frame for statistical analyses. In order to, the following commands are used.

D <- pmApiRequest(query = query, res$total_count, api_key = NULL)

## Documents  200  of  649 
## Documents  400  of  649 
## Documents  600  of  649 
## Documents  649  of  649

M <- pmApi2df(D)

## ================================================================================

M <- convert2df(D, dbsource = "pubmed", format = "api")

## 
## Converting your pubmed collection into a bibliographic dataframe
## 
## ================================================================================
## Done!

Now we use the following commands to get an overview of the data. This information can be gattered in a table suitable for html files. Some attributes like cell positions, cell alignment and so on can be set with different arguments of kable function.

results <- biblioAnalysis(M)

Sometimes, researchers may prefer TO do their analysis in a specific type of document, as only articles. On the other hand, since Rstudio company have been found in 2011, searches are limited to after 2011.

M <- filter(M, M$DT == "JOURNAL ARTICLE" & M$PY >= 2011)
results <- biblioAnalysis(M)

a <- summary(results)

knitr::kable(a$MainInformationDF, caption = "Main information of articles",align = "llccl", format = "html") %>% 
    kable_classic(full_width = F, position = "center")

General information about scientific documents

Here, we can look at some tables and plots which are distracted from data set based on our search strategy.

knitr::kable(a$AnnualProduction, caption = 
           "Annualy Production for scientific Documents",
           align = "cc", format = "html") %>%
           kable_classic(full_width = F, position = "center")

knitr::kable(a$MostProdAuthors, caption = "Top 10 Authors", align = "lclc", format = "html") %>% kable_classic(full_width = F, position = "center")

knitr::kable(a$MostCitedPapers[,1:2], caption = "10 Most Cited Papers",  
             align = "ll",format = "html") %>%  
    kable_classic(full_width = F, position = "center")

knitr::kable(a$MostRelSources, caption = "Top 10 Journals", align =       
        "lc", format = "html") %>% kable_classic(full_width = F, position = "center")

knitr::kable(a$MostRelKeywords, caption = "Top 10 Keywords", align = "lclc", format = "html") %>%
     add_footnote(c("DE: Keywords Extracted from Articles","ID: Keywords
     Extracted from References of Articles"), notation="alphabet") %>%  
     kable_classic(position = "center")

res <- authorProdOverTime(M, k=10)

knitr::kable(res$dfAU[1:3], caption = "Top 10 authors and their timeline,as well annually production, total citations, total citations per year", format = "html") %>% kable_classic(position = "center")

topSO = sourceGrowth(M, top=1, cdf=FALSE)
DF = melt(topSO, id='Year')
ggplot(DF,aes(Year,value, group=variable, color=variable))+geom_line()

topSO = sourceGrowth(M, top=3, cdf=FALSE)
DF = melt(topSO, id='Year')

DF = dominance(results)
knitr::kable(DF, caption = "Some Information about Top 10 Authors", digits = 3, align = "lccccccc", format = "html") %>%
    kable_classic(position = "center")       

knitr::kable(head(sort(table(M$SO_CO),decreasing=TRUE),10), caption = "Top Countries based on Frequency of published articles in Journals", col.names =         c("Country", "Frequency"), align = "lc", format = "html") %>%
        kable_classic(full_width = F, position = "center" )

What does say Lotka’s Law us about these data set?

L=lotka(results)
lotkaTable=cbind(L$AuthorProd[,1],L$AuthorProd[,2],L$AuthorProd[,3],L$fitted)
knitr::kable(lotkaTable, caption = "Frequency Of Authors Based on Lotka's Law", digits = 3, align = "cccc", format = "html",col.names = c("Number of article", "Number of authors", "Frequency based on data", "Frequency based on Lotka's law")) %>%
    kable_classic(full_width = F, position = "center")

Pvalue of two-sample Kolmogorov-Smirnov test between the frequency based on data and the Lotka’s Law is 0.0148932. In significance level of 0.05, this value says us that our data do not follow Lotka’s law.

Collaboration networks for authors

Collaboration network of authors are plotted. As well, the network can be plotted for keywords, universities and countries.

NetMatrix <- biblioNetwork(M, analysis = "collaboration", 
                         network = "authors", sep = ";")
net <- networkPlot(NetMatrix, n = 10, type = "auto", Title = "collaboration Network",labelsize=1, halo = TRUE) 

Thematic map

Thematic Maps are plotted based on (keywords) DE AS follows:

remove.terms.1word = c("aged","map","allergy","demand","rest","workflow","data collection","r","rstudio","data analysis","conservation","review","functional",
    "clinical","identification","data","analysis","network","systematic","r programming","r package","maternal","reproducibility","r language","methods","treatment","r programming language","sars-cov-2","retention","calcium","statistics","open source","quality","methodology","complications","statistical analysis","prognosis","algorithms","software")

synonyms1 <- c("covid-19;coronavirus","gene; genes", "prediction; predicting", "modeling; modelling; resting","emotion; emotional", "adhd; hyperactivity",
      "differentially expressed genes;differentially expressed")
tm1 = thematicMap(M, field = "DE",n.labels = 2, ngrams = 1, remove.terms = remove.terms.1word,synonyms = synonyms1)
plot(tm1$map)

Thematic map is a plot which has been divided to four quadrant: Niche Themes, Motor Themes, Basic Themes and Emerging or declining Themes. For more details refer to (Zhang et al. 2022).

Motor Themes: Quadrant I, located in the upper-right quadrant, named motor
themes, suggested that the themes of the quadrant have developed
and formed important pillars that shape the field of research.

Niche Themes: Quadrant II, located in the upper left quadrant, named niche themes, reflected highly developed but isolated themes.

Emerging or declining Themes: Quadrant III, located in the lower-left
quadrant and named emerging or declining themes, suggested weak development and marginalization of the research field.

Basic Themes: Quadrant IV, located in the lower-right quadrant, was named as basic themes. Although these topics are less developed, they are important to the field of study.

Some diseases (motor themes), like obesity, covid_19, schizophernia, cancer, tuberculosis are discussed well and highly developed and analyzed by Rstudio. on the other side, some statistical and analytical topics such as machine learning, pca (principal component analysis), bibliometrics and bioinformatic analysis.

some diseases (basic themes) like type 2 diabetes mellitus, stroke, differentially expressed genes need to be considered and analyzed more than the present by rstudio, as well as meta analysis, systematic review, network analysis and computational analysis are methods which is reccommedned to use.

Based on this map, there are some themes which have been over discussed (topics covered by niche themes quadrant) in PubMed database. Topics such as, natural language processing, text mining, Pan-Cancer, behavioral science etc. As well, themes in quadrant lll, for example visualization and shiny are of declining themes.

Explanation: some words which we don’t want to be included in the map, as well synonym words are predefined.

Associations among our information

Here, we can see association among Authors, DEs and Journals.

threeFieldsPlot(M)

This plot shows how keywords, authors and journals are related to each other.

Thematic Evolution Plot

Here, We can See Evolution of Topics in RStudio applications field based on DE and TI.

This plot shows themes which have been evolutted during the years.

years=c(2019)

nexus <- thematicEvolution(M,field="DE",years=years,n=100,
          minFreq=3, ngrams = 1,remove.terms = remove.terms.1word,
          synonyms = synonyms1)
plotThematicEvolution(nexus$Nodes,nexus$Edges)

nexus <- thematicEvolution(M,field="TI",years=years,n=100,
          minFreq=3, ngrams = 2,remove.terms = remove.terms.1word,
          synonyms = synonyms1)

## Warning: Expected 2 pieces. Missing pieces filled with `NA` in 1 rows [779].

plotThematicEvolution(nexus$Nodes,nexus$Edges)

References: