environmental-data-science

A scalable computational framework using web data and large language models is presented to quantify organizational environmental actions, demonstrated on U.S. Jewish congregations, comparing keyword retrieval, semantic retrieval, and direct classification approaches.