HedgeSage

A free and open-source hedge fund position aggregator and investment simulator

✨ It uses 13F filings to:

• Aggregate current hedge fund positions
• Backtest the funds' original performance
• Simulate returns that would have been achieved by copying the hedge funds' positions
  • With manager-weighted allocation
  • With manager-weighted allocation and investment value scaled to 100% of the portfolio
• And shows all of this within a simple, neat and minimal UI :)

🚀 Running yourself is very simple:

• Clone the repo.
• Install the dependencies using pip install -r requirements.txt.
• To update live allocations, run python runner_update_allocations.py
• To run the backtests, run python runner_do_backtests.py
• To run the UI live, run cd src_site && npm run dev
• To build the UI in the docs directory, run cd src_site && npm run build

📊 How it works:

• Data for hedge fund live allocations is refreshed daily using a Github Action powered by runner_update_allocations.py, which fetches the hedge funds from the top_funds.json file and runs the refresh function from fetch_hedge_fund_allocations.py. The live allocations are saved to ./sec/allocations, whilst CUSIP conversion cache is saved to ./sec/cusip_conversion_table.json.
• The backtesting entrypoint is runner_do_backtests.py, whilst the backtesting and downloading functionality is implemented in backtester.py and download_data.py, respectively. The backtests are saved to ./sec/backtests, historical allocations are saved to ./sec/past_allocations, whereas price history is saved to ./data/historical.
• The HedgeSage UI is built using Svelte, with shadcn-svelte used as the component library. It acts merely as a front-end to the information stored in this repository. The builds are targeting the docs directory for simplicity in use with GitHub Pages.

TODO:

• Implement fund searching functionality by name/cik
• Filtering based on annualized returns
• Allow to retrieve 13F data from .txt files, letting backtests through years earlier than 2013.
• Allow for inclusion and downloading of non-stock tickers for the backtests
• Unified performance rating
• Allocations last update date
• Saving portfolios with automatic update checking
• Latest allocations update section
• Stock ownership information
• Adjust extreme volatility data sanity checks
• Comparison versus benchmarks
• Incremental backtest updates
• Allocations table sorting/searching
• Search in portfolio creation
• Load/unload cards in the list as the user scrolls (?)
• Politician tracking
• Biggest daily returns of the funds
• Special page for each fund (SEO)
• Portfolio builder
• Fix the live allocations retriever (basically copy functionality from the past allocations retriever)
• Prevent rate limits from sec.gov
• There is an inefficiency in runner_do_backtests.py, as the allocations fetching is done twice
• Make backtests runnable on GitHub Actions
• Allow to export allocations as a CSV
• Add link to repo
• Show when allocations were last changed
• Improve the portfolio builder UX - instead of a drawer allowing you to select portfolios, instead there should be a + button on every one of the hedge fund cards, which will add the portfolio to the portfolio builder, which in of itself could be a drawer containing just the selected portfolios instead. This is better, because that would make better use of the feature-rich cards that the user would look at anyways and avoid the use of the clumsy checkboxes