Insider Screener & Alerts Bot

A Python project that downloads Form 4 and Form 144 SEC filings, filters for high-signal insider trades, cross-references market signals, and supports backtesting.

Features

• Parses and stores insider filings into DuckDB
• Screens for high-value and cluster insider buys
• Cross-references earnings, news, and short interest
• Backtests historical success of trades

Getting Started

pip install -r requirements.txt
python download_filings.py # populate a local database with all filings
python filters.py --min_buy 100000 # search db using filters
python backtest.py --days 30 

Cluster Buys Detected

🏁 Optional Future Add-ons

• UI dashboard (e.g., Streamlit or Dash)
• Insider heatmap by sector
• Sentiment overlay from social media/news
• Broker API alerts (e.g., Alpaca or IBKR paper trading)

Signal Enrichment & Alerts

✅ Goal: Cross-reference with other market data sources

• Earnings calendar (e.g., EODHD, Yahoo Finance, or Nasdaq API)
• Analyst upgrades/downgrades (Finviz scrape, MarketBeat, Barchart)
• Short interest (e.g., via FINRA or third-party APIs)
• News headlines (e.g., RSS feeds, Benzinga, or scraping Yahoo News)
• Flag if insider trade occurred before earnings, after downgrade, or on high short interest

Deliverables:

• enrich_signals.py: joins filing data with enrichment signals
• alerts.py: outputs summary alerts to email/Slack/Telegram

Backtesting Engine**

✅ Goal: Test predictive power of insider trades

• Load historical price data (Alpha Vantage, Polygon, Yahoo Finance, etc.)

• Join filing date with price history:

  • Compute % return after 7, 14, 30, 60 days
  • Label trades as “successful” if return > threshold

• Support backtest filters:

  python backtest.py --filter "officer_buys > 100000 AND cluster = True"

Deliverables:

• backtest.py: returns distribution of returns per filter
• Summary report: hit rate, average return, drawdown

---