PyPoll: A python library automating mining of networks, discussions and polarization on Twitter

Today online social networks have a high impact in our society as more and more people use them for communicating with each other, express their opinions, participating in public discussions, etc. In particular, Twitter is one of the most popular social network platforms people mainly use for political discussions. This attracted the interest of many research studies that analyzed social phenomena on Twitter, by collecting data, analysing communication patterns, and exploring the structure of user networks. While previous works share many common methodologies for data collection and analysis, these are mainly re-implemented every time by researchers in a custom way. In this paper, we introduce PyPoll an open-source Python library that operationalizes common analysis tasks for Twitter discussions. With PyPoll users can perform Twitter graph mining, calculate the polarization index and generate interactive visualizations without needing third-party tools. We believe that PyPoll can help researchers automate their tasks by giving them methods that are easy to use. Also, we demonstrate the use of the library by presenting two use cases; the PyPoll visualization app, an online application for graph visualizing and sharing, and the Political Lighthouse, a Web portal for displaying the polarization in various political topics on Twitter.