This project aims to monitor the lifespan of hyperpartisan content that circulates on Twitter in the period leading up to national elections and democratic consultations. In past research user-generated, hyperpartisan news content has a remarkably short shelf life, which is a marker of the perishable nature of digital content at the center of political debates in liberal democracies. Key to monitoring efforts is the storage of social media data as it is collected, such as relying on public Twitter Streaming API to track the content tweeted by users associated with key global electoral events (US 2020 Election, Brexit Referendum, etc.). The storage resources for the tweets are parsed for real-time archiving of embedded content including images and URLs embedded to the message, hence identifying and archiving the content of webpages tweeted in the context of electoral politics. From the archived data, conducing an analysis workflow will rely on topic models to contrast extant and extinct URL content tweeted in the period leading up to the vote, and during the analysis, an estimation of the size of the retweet cascade that vanished and probe the relationship between content (i.e., hyperpartisan pieces or outright fake news) and content shelf-life.
The goal is to establish metrics for the lifespan of fake news and user-generated, hyperpartisan news articles. The expectation is that the availability of this content will be short-lived and therefore hypothesize that the news cycle of hyperpartisan news deviates from the regular news cycle, including the dynamics of retweet cascades triggered by legitimate news pieces featured on Twitter.