About
CandiData is a collaborative initiative by the Algorithmic Transparency Institute (ATI), the Center for Tech and Civic Life, the Media and Democracy Data Cooperative, the Social Media Archive at the University of Michigan, and the Political Science Department at the University of Kentucky. Dedicated to documenting political speech, its 2024 US elections mission focused on creating a structured repository of candidates’ social media communications, offering accurate data across platforms such as Facebook, TikTok, Threads, and Rumble.
By compiling datasets of social media accounts and activity of US federal and state candidates for the 2024 elections, CandiData bridges a critical gap between unstructured online political discourse and researchers seeking reliable data for analysis
Challenge
Political candidates and campaigns rely heavily on social media for voter engagement, fundraising, and narrative control. However, researchers face persistent hurdles in accessing comprehensive data due to restrictive platform policies, API cost barriers, and scattered data points. To address these challenges, CandiData integrated multichannel data into Junkipedia, a project by ATI and supported by The Bright Initiative, where they are able to add critical real-time monitoring capabilities.
Impact
CandiData developed a detailed database of candidate social media accounts, now publicly available on platforms like SOMAR and GitHub. The project also includes interactive dashboards that support real-time monitoring of candidate messaging, offering insights into trends and communication patterns. These resources provide researchers, journalists, and civic organizations with valuable tools to analyze political speech and its influence on public discourse.
