A personal fantasy baseball analysis app built on the Yahoo Fantasy API. Blends Statcast data, consensus projections (Steamer/ZiPS/ATC), and AI-powered analysis into a single dashboard for roster optimization, trade evaluation, waiver recommendations, and weekly matchup breakdowns.
Built by Brian Renshaw with Claude Code.
How to use every feature in the app — dashboard, roster optimizer, trade analyzer, waiver wire, stats explorer, projections, matchups, and the AI assistant. Includes scoring breakdowns, stat definitions, and a full glossary.
How the app is built — FastAPI services, SQLAlchemy models, ETL pipeline, projection blending, and the data flow from Yahoo/Statcast/FanGraphs into the database. Start here if you want to understand or extend the code.
Everything AI powers in this app — from overnight podcast transcription and expert content analysis to personalized daily briefings, smart lineup recommendations, and a conversational assistant that knows your league.
The content ingestion system that pulls from fantasy baseball blogs and podcasts, then uses Claude to generate daily and weekly analysis reports with league-specific context. See example newsletter below.
- Dashboard — League standings, weekly matchup projections, and roster overview
- Roster Optimizer — ILP-based daily/weekly lineup optimization (PuLP)
- Trade Analyzer — VORP + z-score trade values with side-by-side comparison
- Waiver Wire — Composite scoring with Steamer ROS projections and Statcast signals
- Stats Explorer — Interactive Plotly charts across batting, pitching, and Statcast metrics
- Player Comparison — Head-to-head stat comparison with radar charts
- Projections — Blended consensus projections with buy/sell signal detection
- Intel Reports — AI-generated daily analysis from ingested blogs and podcasts
- AI Assistant — In-app Claude-powered chat with league context
- Matchup Quality — Park factors, platoon splits, and opposing pitcher adjustments
- Projection Accuracy — Weekly tracking of projection accuracy vs actuals
| Layer | Tools |
|---|---|
| Backend | FastAPI, SQLAlchemy (async), APScheduler |
| Frontend | HTMX, Tailwind CSS, Plotly.js, Jinja2 |
| Database | SQLite via aiosqlite |
| Data | Yahoo Fantasy API (yfpy), pybaseball, MLB-StatsAPI |
| AI | Anthropic Claude API |
| Optimization | PuLP (Integer Linear Programming) |
| Deployment | Docker, Fly.io |
-
Install uv
-
Configure environment
cp .env.example .env # Edit .env with your Yahoo API credentials and Anthropic API key -
Install dependencies and run
uv sync uv run uvicorn app.main:app --reload --port 8000
-
First run — The app will trigger a browser-based Yahoo OAuth flow. After authenticating, tokens refresh automatically.
See the App Manual for a full walkthrough and the Architecture guide for details.
uv run ruff check . # Lint
uv run ruff format . # Format
uv run pytest # Test
uv run python -m app.etl.pipeline # Run ETL manually
uv run python -m scripts.daily_analysis # Generate daily intel reportsapp/
main.py # FastAPI entry point + auth middleware
config.py # pydantic-settings configuration
database.py # SQLAlchemy async engine
models/ # ORM models (players, stats, projections, rosters, etc.)
services/ # Business logic (Yahoo, stats, trades, waivers, optimizer, AI)
routes/ # FastAPI route handlers
templates/ # Jinja2 HTML templates
etl/ # Data pipeline (extract, transform, load)
static/ # CSS + JS (charts, tables, tooltips)
scripts/ # Backtesting, data pipeline, content ingestion
docs/ # User guide, methodology, pipeline docs
tests/ # pytest test suite
The app's content pipeline ingests fantasy baseball podcasts and blog articles, then uses Claude to generate a personalized weekly intel report. Each report includes roster-specific analysis with sentiment ratings, lineup recommendations, injury alerts, trade signals, waiver targets, and league-contextualized advice — all citing the original sources.
Preview pages (click to expand)
