Benchmark Datasets for Table Discovery Using Natural Language Questions

Dataset Description Tables Questions
Public BI Used for business intelligence and data warehousing tasks. It contains data typically found in reports, dashboards, and business decision-making systems Download Download
Chicago Open Contains publicly available data from the City of Chicago, representing various aspects of civic operations, including transportation, crime, and public services Download Download
Chembl A scientific database containing bioactivity data on drug-like molecules, used in pharmaceutical research Download Download
FetaQA Collected from Wikipedia, designed for factual table question answering (QA) tasks, where users pose questions to retrieve factual information from tables Download Download
Adventure Works Simulates an enterprise database, with tables related to business operations, sales, and manufacturing Download Download