Prediction Market Analysis is a framework for collecting and analyzing prediction market data that provides the largest publicly available dataset of Polymarket and Kalshi market and trade information.
The framework addresses the challenge of accessing and studying prediction market microstructure by bundling pre-collected datasets alongside tools for gathering new data and running statistical analyses. It collects market metadata, trade history via APIs and blockchain sources, and stores everything in Parquet format with automatic progress tracking. The approach separates data collection, storage, and analysis into distinct phases, allowing researchers to work with historical data or extend the dataset with fresh information.
Researchers studying prediction markets, particularly those analyzing Polymarket and Kalshi, should consider this tool if they need structured access to trade-level data without building collection infrastructure from scratch. The pre-collected dataset eliminates the initial barrier of bootstrapping data collection. The framework is extensible through custom analysis scripts, making it suitable for projects that require both standard statistical outputs and domain-specific investigations. Python 3.9 or later is required.
The project shows active maintenance with regular updates to data collection indexers and analysis capabilities. Documentation is provided for both data schemas and the process of writing new analyses. The maintainer responds to issues and accepts pull requests, indicating ongoing engagement with the user base.
