Buffett-Style Stock Dashboard

An implementation of the book "Warren Buffett and the Interpretation of Financial Statements: The Search for the Company with a Durable Competitive Advantage". This app evaluate stocks using a simplified Warren Buffett methodology over a 5‑year window.
It is built with Streamlit, uses on-demand live prices, and persists fundamentals in Parquet files (no SQL, no Streamlit caches).

✨ Features

• Portfolio view with manual or on-demand live prices (via Financial Modeling Prep).
• Stock Lookup with a 5-rule Buffett-style scorecard and Best Entry Points (P/B, Graham Net-Net, P/E relative).
• Comparisons: multi-line charts across holdings (ROE ≥ 15%, D/E ≤ 50%, Equity growth, Profit growth, FCF ≥ 0).
• Parquet persistence: .cache/fundamentals/<TICKER>.parquet for fast repeat lookups.
• Currency-aware labels for common suffixes (e.g., .SR → SAR).

Note: This tool is for educational purposes and not investment advice.

🚀 Quickstart

1) Requirements

• Python 3.10+ recommended

2) Create a virtual environment & install

python -m venv .venv
source .venv/bin/activate  # Windows: .venv\Scripts\activate
pip install -r requirements.txt

3) Provide your FMP API key

You can paste it in the sidebar input when the app runs, or set it in your shell:

export FMP_API_KEY=your_key_here      # PowerShell: $Env:FMP_API_KEY="your_key_here"

Alternatively, for local dev you can store it in .streamlit/secrets.toml (this file is gitignored):

# .streamlit/secrets.toml
FMP_API_KEY = "your_key_here"

4) Run

streamlit run streamlit_app.py

📂 Project structure (key files)

.
├── streamlit_app.py               # Main app
├── data_providers.py              # HTTP fetchers (with timeout & retries)
├── parquet_store.py               # Parquet persistence (load/upsert per ticker)
├── buffett_eval/
│   ├── __init__.py
│   └── metrics.py                 # 5-rule scorecard logic
├── data/
│   ├── sample_portfolio.xlsx      # Example portfolio (no live prices)
│   └── sample_fundamentals.csv    # Example fundamentals
├── .streamlit/
│   └── config.toml                # Theme
├── requirements.txt
├── README.md
└── .gitignore

🧾 Data formats

Portfolio Excel (minimal columns)

Column	Type	Notes
Ticker	string	e.g., AAPL, MSFT, 2222.SR
Shares	number
AvgCost	number	average cost per share

Optional: Company, Sector, etc.

Fundamentals CSV (per row per year)

Required columns (case-sensitive):

ticker, company, year, revenue, net_income, shareholders_equity,
total_debt, shares_outstanding, free_cash_flow, current_assets, total_liabilities

🧠 Buffett-style rules (5y window)

• Equity growing (CAGR > 0)
• Debt-to-Equity < 50% (latest year)
• Profit (Net Income) growing (CAGR > 0)
• ROE ≥ 15% in ≥ 4 of 5 years
• FCF positive in all 5 years

Best Entry Points (not in score):

• P/B: buy if Price ≤ 0.8 × BVPS (BVPS = Equity / Shares).
• Graham Net-Net: buy if Price ≤ ⅔ × NCAV/share (NCAV = Current Assets − Total Liabilities).
• P/E (relative): buy if Current P/E ≤ 0.70 × Reference P/E (industry preferred; fallback to company 5y median P/E).

💾 Caching

• Fundamentals are stored under .cache/fundamentals/<TICKER>.parquet.
• The app loads from uploaded CSV, then Parquet cache, then API (and saves back to Parquet).

🌍 Currencies

Basic heuristics: .SR→SAR, .L→GBP, .TO→CAD, .HK→HKD, .F/.DE→EUR; else default USD.
You can extend detect_currency() in streamlit_app.py for more exchanges.

🛠 Troubleshooting

• No data / Missing NCAV: Ensure your fundamentals CSV includes current_assets & total_liabilities, or supply an FMP key.
• Rate limits / timeouts: The app uses timeouts and light retries; try again or throttle lookups.
• Mixed currencies: Portfolio totals are disabled for mixed-ccy portfolios (FX conversion not yet implemented).
• Charts show NaN/inf: Usually equity is zero/negative; the app guards divisions, but verify your data.

🔧 Customization

• Change thresholds (ROE 15%, D/E 50%) in metrics.py and chart reference lines in streamlit_app.py.
• Add a Cache Manager panel to list/delete Parquet files if desired.

📄 License

Choose what fits your needs (MIT/Apache-2.0/BSD-3-Clause).

— Generated on 2025-08-24