ABSTRACT
When a team of developers are creating new software, they most
likely will use libraries of code that can assist in a given required
feature. One source to find these libraries can be popular questionanswer
websites, blogs, personal web pages and the usage of tools
that can automatically suggest libraries. Popularity might be one
criterion that developers can use when choosing a library. In this
work, we performed an empirical evaluation through mining Python
projects hosted in GitHub to identify the most popular used modules.
We selected 129 projects based on specific criteria, one of them
being the number of stars that reflects their popularity. To automate
the data extraction process, we developed the PySniffer, an opensource
tool that performs a static code analysis in Python scripts,
checking which modules from both the standard library and external
modules are used in a project. Our tool also has a front-end that
can display the data more friendly with statistical information. As
a result, we generated a list with the top used modules in Python
projects hosted in GitHub, serving as complementary information
alongside the most popular libraries informed in personal blogs and
websites.