AI 驅動的美股投資研究平台 — 基本面分析、SEC 文件、內線交易追蹤、技術分析，全部繁體中文。

Live Site: yennanliu.github.io/finance_data

A research platform that combines SEC filings, AI-generated analysis, and automated pipelines to produce investment research reports for US equities — all in Traditional Chinese.

Daily Analysis (25 tickers, fundamental + technical): 0050, 2330.TW, TSLA, PL, GRAB, TSM, GOOG, AMZN, MSFT, SOFI, PLTR, RKLB, ONDS, AVAV, KTOS, META, AMD, NVDA, NU, VST, ORCL, INTC, SPCX, AVGO, NBIS

Analysis Types (12): fundamental-analysis, technical-analysis, stock-eval, stock-valuation, financial-report-analyst, earnings-call-analysis, insider-trading, institutional-ownership, sector-analysis, economics-analysis, portfolio-review, report-generator

SEC 10-K Filings: 30+ companies across tech, defense, energy, and financials

AI Research Notes: ONDS, RKLB, AVAV, RCAT, TSLA, NEE, AMZN

• Multi-provider LLM support — generate with Claude, OpenAI, or Google Gemini (Gemini gemini-2.5-flash is now the daily-CI default); see scripts/analysis/config/providers.py
• Site size cut 3.1 GB → 503 MB (-84%) — search-index trimming, pruned navigation, WebP screenshots, GitHub-raw PDF links, 120-day retention, and a single nightly deploy. Full write-up: 部署效能調校全紀錄
• Responsive content — tables, charts, and articles render cleanly across iOS / Android / web
• Advanced analysis pipeline — earnings-call, insider-trading, institutional-ownership, and interactive HTML reports via advanced_analysis.yml
• Automated QA & housekeeping — daily report-quality checks (qa_report_quality.yml), refusal-post cleanup (cleanup_refusals.yml), and daily progress summaries (daily_progress.yml)
• Taiwan market coverage — added TW-listed tickers (0050, 2330.TW)

┌─────────────┐     ┌──────────────────┐     ┌────────────────┐     ┌──────────────┐
│  Data Fetch  │────▶│   AI Analysis    │────▶│  Auto Deploy   │────▶│  Browse Site  │
│  yfinance    │     │ Claude / OpenAI  │     │  GitHub Actions │     │  GitHub Pages │
│  SEC EDGAR   │     │ Gemini · 繁中報告 │     │  Nightly CI/CD  │     │  MkDocs       │
└─────────────┘     └──────────────────┘     └────────────────┘     └──────────────┘

1. Data Fetch — yfinance API pulls live financial data; scripts download filings from SEC EDGAR
2. AI Analysis — scripts/generate_analysis.py sends data to your chosen provider (Claude, OpenAI, or Gemini), which produces structured Markdown reports with ASCII charts
3. Auto Deploy — GitHub Actions runs analysis daily and publishes via a single nightly mkdocs build (consolidated from 42 per-job deploys)
4. Browse Site — Reports published at yennanliu.github.io/finance_data with search, dark/light mode, responsive layout, and category navigation

pip install -e ".[dev]"   # installs anthropic, openai, google-genai, yfinance, mkdocs, …

# Set the API key for the provider you want to use
export ANTHROPIC_API_KEY="sk-..."   # Claude
export OPENAI_API_KEY="sk-..."      # OpenAI
export GEMINI_API_KEY="..."         # Gemini (daily-CI default)

# Generate a fundamental analysis for AAPL (default provider: Claude)
python3 scripts/generate_analysis.py AAPL

# Choose provider and analysis type
python3 scripts/generate_analysis.py MSFT --analysis-type technical-analysis --provider gemini
python3 scripts/generate_analysis.py NVDA --analysis-type stock-valuation --provider openai

Output saved to ai_gen_report/stock/aapl/fundamental_analysis_YYYY-MM-DD.md

pip install requests

# Download 5 most recent 10-K reports for Apple
python3 scripts/download_10k.py AAPL

# Download for multiple companies
python3 scripts/download_10k.py AAPL MSFT TSLA

# Download 10-K PDFs
python3 scripts/download_10k_pdf.py meta-platforms-inc

pip install mkdocs-material mkdocs-awesome-pages-plugin mkdocs-minify-plugin

python3 scripts/build_docs.py
mkdocs serve

finance_data/
├── ai_gen_report/
│   ├── stock/        # AI-generated analysis reports (Markdown)
│   └── market_news/  # Daily AI-curated market news
├── notebook_llm/     # NotebookLM deep research notes
├── 10-k/             # SEC 10-K annual reports (PDF)
├── 10-q/             # SEC 10-Q quarterly reports
├── 13-f/             # SEC 13-F institutional holdings
├── 6-k/              # SEC 6-K foreign issuer reports
├── investor_day/     # Investor day presentations
├── scripts/          # Automation tools
│   ├── generate_analysis.py   # AI analysis generator
│   ├── build_docs.py          # MkDocs content builder
│   ├── download_10k.py        # SEC EDGAR downloader
│   └── download_10k_pdf.py    # PDF downloader
├── docs/             # MkDocs source (auto-generated)
├── .github/workflows/
│   ├── deploy.yml             # Build & deploy GitHub Pages
│   ├── daily_analysis.yml     # Daily AI analysis cron
│   └── daily_market_news.yml  # Daily market news cron
└── mkdocs.yml        # MkDocs Material configuration

• Financial Data: Yahoo Finance via yfinance (no API key needed)
• SEC Filings: SEC EDGAR
• AI Analysis: Claude (Anthropic), OpenAI, and Gemini — provider is selectable per run
• Annual Reports: annualreports.com
• Google Finance - 10K Reports: Google Finance

• 把站台從 3.1GB 砍到 503MB：finance_data 部署效能調校全紀錄 — engineering write-up on the MkDocs site-size and deploy-performance tuning (search index, navigation pruning, WebP, retention, nightly deploy): yennj12.js.org

All analysis is for educational and research purposes only. Nothing on this site or in this repository constitutes investment advice. Invest at your own risk.