Analyses price performance and correlation between semiconductor and tech stocks (NVDA, MU, MSFT) using Python. Produces publication-quality charts and a summary statistics table from 2 years of daily market data.
- Python 3.13
- yfinance — market data ingestion
- pandas / numpy — data cleaning and analysis
- matplotlib — static two-panel performance chart
- plotly — interactive cumulative returns chart
ai-market-engine/
├── config.py # Single source of truth: tickers, date range
├── fetch.py # Download CSVs from yfinance (run once)
├── main.py # Pipeline entry point
├── src/
│ ├── data_cleaning.py # clean, normalize, flag outliers
│ ├── analysis.py # returns, volatility, summary stats
│ └── visualization.py # two-panel matplotlib chart
└── data/ # Downloaded CSVs (git-ignored)
python3 -m venv .venv
source .venv/bin/activate
pip install -r requirements.txtpython -m venv .venv
.venv\Scripts\activate
pip install -r requirements.txtStep 1 — Download data (run once):
python fetch.pyStep 2 — Run the full pipeline:
python main.pyOutputs:
charts/performance_overview.png— static two-panel chartcharts/interactive.html— interactive cumulative returns chart- Summary statistics printed to terminal
| Mean Daily Return | Annualized Vol | Max Drawdown | Total Return | |
|---|---|---|---|---|
| NVDA | +0.32% | 51.1% | -36.9% | +287.4% |
| MU | +0.32% | 57.4% | -57.6% | +249.2% |
| MSFT | +0.07% | 22.2% | -23.7% | +32.3% |
Parkinson vs Close-to-Close Volatility
| Close-to-Close Vol | Parkinson Vol | Ratio | |
|---|---|---|---|
| NVDA | 3.06% | 2.44% | 0.80 |
| MU | 3.45% | 2.50% | 0.73 |
| MSFT | 1.33% | 1.11% | 0.84 |
Parkinson volatility (using daily High/Low range) consistently estimates lower vol than Close-to-Close std. The gap measures overnight gap risk — price jumps between yesterday's close and today's open that are invisible to intraday vol estimators.
MU has the lowest ratio (0.73), meaning ~27% of its measured volatility comes from overnight gaps driven by after-hours memory pricing updates and earnings surprises. MSFT's higher ratio (0.84) reflects mega-cap stability with fewer overnight dislocations.
Implication: For overnight hedging strategies, Close-to-Close vol is the appropriate risk measure. For intraday strategies, Parkinson is more accurate.