Analysis of CRAN packages

This repo aims to analyse some caracteristics of CRAN package.

Data collecting

The data collecting part was inspired by this question on SO.

We collect the datas :

• through tools::CRAN_package_db() to get the descriptions ;
• through scraping from CRAN, with polite and rvest for the views;
• through cranlogs to get the download stats.

Some inspiration for the cleaning come from the grapher package (and this function).

Some possible subjects

• by how many people and by who are packages built and maintained ?
• same for organizations ?
• analysis of collaboration on packages through author list ;
• identify "core" packages (the one that people download themselves) and underlaying packages (the one downloaded by packages) ;
• GIF of the growing of CRAN ;
• do packages have nationality ?
• predict the topic of packages with depends and description and compare it to in views subject .

Resources

Some work as be done on the subject and could be useful :

• https://arxiv.org/pdf/2007.07820.pdf
• https://blog.revolutionanalytics.com/2015/08/differences-in-the-network-structure-of-cran-and-bioconductor.html?utm_content=bufferabeb1&utm_medium=social&utm_source=twitter.com&utm_campaign=buffer
• https://shiny.john-coene.com/cran/
• https://cran.r-project.org/web/packages/pkggraph/vignettes/vignette_pkggraph.html