Measure how markets digest scheduled events (CPI, FOMC, earnings) using free Yahoo Finance data.

• Pulls prices with yfinance-pl (Polars-native).
• Defines pre/post event windows.
• Computes pre/post returns, realized vol changes, max drawdowns, and reaction speed (minutes to peak move).
• Compares reactions across assets to see which markets price news fastest.

uv run event-impact --assets "SPY,QQQ,GLD,TLT,EURUSD=X,CL=F" --interval 1h --categories "cpi,fomc" --year 2025 --pre-hours 24 --post-hours 24

• Append local files: --events-file my_events.csv (columns: name,category,timestamp[,tickers]) or JSON array with the same keys.
• Append ICS feeds: --events-ics-url https://.../calendar.ics [--events-ics-category macro] (expects DTSTART/SUMMARY).
• Append FRED release dates (requires FRED_API_KEY): --events-fred-release-ids "cpi=9,fomc=10,employment=50" [--fred-start YYYY-MM-DD --fred-end YYYY-MM-DD]. Dates are stamped at 08:30 ET by default.
• Built-in calendars: CPI and FOMC schedules for 2024 and 2025 plus a small earnings sample. Default year is 2025; override with --year 2024 if needed.

• BLS/BEA/other agencies often publish ICS calendars; pass the URL via --events-ics-url.
• FRED release dates via API using your key (release ids documented on FRED; e.g., CPI=9, FOMC Press Release=10, Employment Situation=50, GDP=53).
• For earnings, drop your own CSV/JSON from broker feeds/FTP exports.

Example with FRED release dates:

uv run event-impact \                                             ─╯
  --assets "SPY,QQQ,GLD,TLT,EURUSD=X,CL=F" \
  --interval 1h \
  --categories "cpi,fomc" \
  --year 2025 \
  --events-fred-release-ids "cpi=9,fomc=10" \
  --fred-start 2025-01-01 --fred-end 2025-12-31 \
  --pre-hours 24 --post-hours 24 \
  --output-csv data/impacts_2025.csv

• Events are a built-in 2024/2025 calendar for CPI (8:30am ET), FOMC (2:00pm ET statements), and a small earnings sample. Extend by adding your own events via the flags above.
• For intraday studies, stick to 1h interval to avoid API throttling; 1d works for longer windows.
• Reaction minutes are relative to the event timestamp, using the first time the absolute cumulative return peaks after the announcement.

This project in any shape or form is purely educational and should not be treated as financial advice.