As part of my dissertation, I scraped over 300,000 online video and audio sermons from congregations across the United States. I utilized Google API, YouTube API, and Large Language Models (LLMs) to scrape the content of sermons and analyze their content for political messages on climate change, abortion, immigration, and other sociopolitical issues of our modern times.

My dataset includes geolocation of over 3,000 congregations, paired with demographic and ideological variables of their surrounding neighborhood (based on Census and voting data) in addition to the scraped sermon text. I can be reached at nela.mrchkovska@wu.ac.at for any questions about the data.