The Unwind: the scripts behind the data

Put the scripts in one folder with the input files they name, and run them in this order with Python 3
(pandas 2.3.3, pyarrow 21.0.0: pip install -r requirements.txt). Each step writes the next step's input.

1. python3 screen_build.py        every 13F filer for 30 June and its holdings -> screen_2026Q2.csv, holdings_2026Q2.csv.gz
2. python3 hf_universe.py         the managers that run hedge funds -> hf_universe.csv
3. python3 figi_map.py; python3 name_map.py   each security to a ticker -> cusip_ticker.csv
4. python3 prices_pull.py; python3 prices_volume_pull.py   daily prices and volume -> prices_3y.csv, volume_daily.csv
5. python3 run_full.py            each 30 June book priced at 16 October, scored on its filings -> full_scores.csv, forced_sales.csv
6. python3 avg_leverage.py        the equity funds at the SEC's average leverage -> avg_leverage_scores.csv
7. python3 cascade_liq.py         each fund's required sale spread across its holdings -> cascade_liq_stocks.csv, cascade_liq_funds.csv
8. python3 adv_schedule_d.py      each fund's prime brokers from Form ADV Schedule D -> adv_private_funds.csv

The original sources (SEC 13F data sets, Form ADV, SEC Private Funds Statistics, Yahoo prices) are listed,
with links, on the site's Data archive page, section 2.
