Sector Rotation Screen

A weekly methodology demo. Open source, transparent rules.

Every Sunday afternoon, GitHub Actions runs an 11-SPDR-ETF screen against three signals — seasonality, economic-cycle fit, and relative strength vs SPY — runs a 15-year backtest, and asks Claude for a brief commentary on what the output said.

⚠️ Not financial advice. This is auto-generated each week by Claude (Anthropic's AI model). Brian Beals is not a registered investment advisor, and Claude is not licensed to provide personalized financial advice. The screener is a research and methodology demo, not a recommendation system. Past performance does not predict future results. Do your own research before making any investment decisions.

This week — May 10, 2026

Dashboard ↗

Score table, seasonality heatmap, RS bars, equity curve, cycle context.

Commentary ↗

Claude's plain-language reading of this week's screen.

Why this exists

Sector rotation isn't an investment thesis I'm pitching. It's a testbed for the kind of analytical workflow I'd build for a CIO or CDO who wants to evaluate macro exposures across business units: ingest data, score against multiple signals, output something humans can read, backtest the methodology before trusting it.

The same pattern works for portfolio companies, supply-chain risk, customer-segment health, or any question where "let's score it against multiple signals and see how that would have played out historically" beats gut feel.

Open source

The screener, methodology, weekly automation, and license live at github.com/brianbeals/sector-rotation-screener.

Past runs

Browse the date-stamped archive →