gao_reports
One row per U.S. Government Accountability Office (GAO) product, ingested from the public GAO reports RSS feed by build_gao_reports. The federal oversight layer over the rulemakings this dataset tracks — GAO's audits, evaluations, and recommendations on how agencies implement laws and rules. GAO's bulk/search surfaces are bot-blocked, so the RSS feed (a ~25-item recent-products window) is the only anonymous machine-readable source; the table is therefore an append-only accumulator that grows into a rolling history over successive daily runs. Deduped on report_id. All columns are stored as VARCHAR.
- Parquet file:
gao_reports.parquet - Queryable via MCP
query_sql: Yes
| Column | Type | Description |
|---|---|---|
report_id |
VARCHAR |
GAO product id parsed from the product URL (e.g. gao-26-107974), lowercased. Primary key / dedup key. |
title |
VARCHAR |
Product title as published in the feed. |
report_type |
VARCHAR |
GAO product type. The reports feed does not tag a finer type, so this defaults to Report. |
published_date |
VARCHAR |
Publication date parsed from the feed's pubDate to an ISO date string (e.g. 2026-07-17). Sort key. |
abstract |
VARCHAR |
The product summary from the feed (GAO's What GAO Found / Why GAO Did This Study narrative). |
agencies_json |
VARCHAR |
Reserved JSON array of agencies the product covers. The RSS feed carries no structured agency tags, so this is [] pending a future enrichment source. |
topics_json |
VARCHAR |
Reserved JSON array of topics/subjects. The RSS feed carries no structured topic tags, so this is [] pending a future enrichment source. |
url |
VARCHAR |
Canonical gao.gov URL for the product. |