Free, open, and all ten years at once.
All ten years ship in one release, on one day. There is no monthly schedule. Inside the release the files are organized by month (120 folders), with a SQLite database of everything and Parquet by table, under the gametidedata account everywhere. Around the raw files, fifteen ways in: for Excel and Power BI, for SQL and Python, and for AI agents. Each one opens on release or soon after. Sign up and you will hear when it does.
Release: v1.0 (in final build)
Download
The whole dataset is published at once. It is organized by month so you can take what you need: one folder per month-end date, with the date in every file name, zipped by month, plus the SQLite file with everything in one place. Every route below opens on release day; sign up and you will hear the moment it does.
GitHub Releases
The full release: a zip for each of the 120 month folders, the release notes and the verification package.
Sign up for releaseHugging Face
One dataset per table, with Parquet files, and a one-line loader: load_dataset("gametidedata/..."). The best route for the large SQLite file.
Sign up for releaseUse it in the tools you already have
For finance and FP&A people. No code, no installs.
Stable file links
A fixed address for every monthly file, for Excel's Power Query, Google Sheets and pandas. The addresses never change between releases.
Sign up for releaseExcel workbook
A workbook with the Power Query connections prebuilt: pick a month, refresh, and the ledger, AR aging and ARR roll-forward load.
Sign up for releasePower BI template
A .pbit with the data model, relationships and the headline measures already defined, so you start at the visuals.
Sign up for releaseOData feed
Power BI and Excel connect to OData natively. The same tables as the API, reachable from the Get Data menu.
Sign up for releaseQuery it
For SQL and Python people, and for anyone curious.
SQL playground in the browser
Run SQL against the full dataset on this site, with nothing installed. DuckDB runs in your browser and reads the Parquet files directly.
Sign up for releaseSQLite database
All 120 months in one file. Open it in DB Browser, DBeaver, Python or any tool that speaks SQLite.
Sign up for releaseStarter notebooks
Kaggle and Google Colab notebooks that load one month and walk through one analysis, so nobody starts from a blank page.
Sign up for releasePython package
pip install gametide: downloads, caches and returns any table as a DataFrame, by month and system.
Sign up for releaseBuild with AI
For people building agents, and for people who want to ask the data questions in plain language.
AI pack
A plain-text pack (schema summary, metric definitions, the story in brief) written to be pasted into Claude, ChatGPT or Copilot so you can start asking questions about the data.
Sign up for releaseMCP server
AI agents query each system the way they would query the real tool, with its own IDs and field names, and practice combining several systems. Free keys with manual approval.
Sign up for releaseRead-only REST API
Every table, filterable by month and key, with an OpenAPI page. Reads the same SQLite the generator writes. Free keys with manual approval.
Sign up for releaseSystem-shaped endpoints
Endpoints that follow the shape of the real Salesforce and Sage Intacct APIs, so you can practice real integrations against a company that cannot mind.
Sign up for releaseGet notified at release
One email when the data ships, and occasional notes when something new arrives. Nothing else, and you can leave any time.
License
The data is published under the Creative Commons Attribution 4.0 license (CC BY 4.0). Anyone may use it, share it and build on it, including commercially, with credit to GameTide. A suggested credit: "GameTide synthetic company data, gametidedata.org, CC BY 4.0."