Bulk History in Parquet: Congress, Insider and 13F Data for Backtesting

Nightly full-history files of every Bourse dataset in Parquet, CSV and NDJSON, delivered through pre-signed Amazon S3 links, so quants can backtest before going live.

Nobody plugs a live feed into a strategy before running it through years of history. Every night Bourse writes each dataset to whole-history files built for that.

Key takeaways

• Formats: Parquet (Snappy compression, 250,000 rows per part), gzipped CSV and gzipped NDJSON, with the same field names as the API, including timestamps and identifiers.
• API Pro gets CSV and NDJSON. Enterprise ($1,000 a month) also gets Parquet, served from Amazon S3.
• GET /v1/exports returns pre-signed URLs valid for one hour; files refresh every night.
• Congress trades go back to 2012 in the current data; 13F holdings of tracked investors span many years of filings.

Example

`` import bourse df = bourse.download_bulk("congress", format="parquet", read=True) # Enterprise paths = bourse.download_bulk("insiders", format="csv.gz") # API Pro ``

Parquet loads into pandas or Polars much faster than CSV and takes a fraction of the space, because it stores typed columns and compresses them.

Point-in-time

Use processed_by_bourse_timestamp and filed_at_source_timestamp in the files to replay history exactly as it became public. See point-in-time data and the developer page.

Bourse provides financial data and editorial information, parts of which are generated by AI. This page is for general education only. It is not personalised investment advice or a recommendation to buy or sell any financial instrument. Make your own decisions.