stock-market-screener 0.1.2

A data-driven stock screener that analyzes Dow 30 stocks using technical indicators

Stock Market Screener

A data-driven stock screener that analyzes Dow 30 stocks using technical indicators and machine learning-inspired scoring algorithms to identify promising investment opportunities.

Overview

This project evaluates 30 major stocks from the Dow Jones Industrial Average based on multiple technical indicators and combines them into a comprehensive scoring system. The screener uses recent market data (from 2025-01-01 onwards) to calculate momentum, trend, and volume signals.

Quick Start

If you're in a hurry:

1. Clone/download the repo
2. Create virtual environment: python -m venv .venv && .venv\Scripts\activate (Windows) or python3 -m venv .venv && source .venv/bin/activate (Mac/Linux)
3. Install packages: pip install -r requirements.txt
4. Run screener: python stock_screener.py

That's it! Results will display in your terminal and be saved as CSV + PNG charts.

Features

• Multi-Factor Analysis: Combines three complementary scoring metrics:

  • MA50 Score: Moving average convergence/divergence analysis using 50-day and 200-day moving averages
  • RSI Momentum Score: Relative Strength Index momentum detection to identify overbought/oversold conditions
  • Volume Score: Volume analysis combined with price trend confirmation

• Normalized Scoring System: All metrics are normalized between -1 and +1 for easy comparison

• Final Composite Score: Averaging of all three components for a holistic view

• Visual Analytics:

  • Bar chart visualization of final scores by ticker
  • Color-coded heatmap for quick pattern recognition (red = negative, green = positive)

Data Source

• API: Yahoo Finance (via yfinance)
• Stocks: Dow 30 constituents (30 large-cap US stocks)
• Timeframe: 1-year rolling analysis (Jan 2025 - Jan 2026)
• Frequency: Daily OHLCV data with auto-adjustment for splits and dividends

Technical Stack

• Data Analysis: Pandas, NumPy
• Technical Indicators: TA-Lib (TALib)
• Visualization: Matplotlib, Seaborn
• Finance Data: yfinance

Installation

1. Clone or Download the Repository

Download this project or clone it from the repository link.

2. Create a Virtual Environment (Recommended)

On Windows:

python -m venv .venv
.venv\Scripts\activate

On macOS/Linux:

python3 -m venv .venv
source .venv/bin/activate

3. Install Dependencies

pip install -r requirements.txt

Note on TA-Lib: If you encounter issues installing TA-Lib, refer to the TA-Lib installation guide. It requires a compiled C library.

Usage

You can run the stock screener in two ways:

Option 1: Run the Python Script (Easiest)

python stock_screener.py

This will:

• Download historical data for all 30 Dow stocks
• Calculate scores for all stocks
• Display results in the terminal
• Generate and display visualizations
• Save results to a timestamped CSV file
• Save chart images (bar_chart.png and heatmap.png)

Option 2: Run the Jupyter Notebook

jupyter notebook "Stock Market Screener.ipynb"

Then run all cells in order. The notebook provides the same functionality with interactive exploration capabilities.

Scoring Methodology

MA50 Score

Combines three factors:

• Distance from MA50 (50%): How far the current price is from its 50-day moving average
• MA50 Slope (30%): The direction and strength of recent moving average trend
• Trend Regime (20%): Whether price is above/below the 200-day moving average

RSI Momentum Score

Measures the rate of change of RSI:

• Normalizes recent RSI changes by rolling volatility
• Positive values = RSI strengthening (bullish momentum)
• Negative values = RSI weakening (bearish momentum)

Volume Score

Combines volume and price action:

• Volume Component (50%): Current volume relative to 20-day average
• Trend Confirmation (50%): Whether volume increases on up or down days

Output

The screener produces three types of output:

1. Terminal Output

• Displays all stock scores in a formatted table
• Shows progress as data is downloaded and calculated

2. CSV Report

• Automatically saved as stock_screener_results_YYYYMMDD_HHMMSS.csv
• Contains all individual scores and final composite scores
• Can be imported into Excel or other tools for further analysis

3. Visualizations

• Bar Chart (stock_screener_bar_chart.png): Final scores ranked by ticker for easy comparison
• Heatmap (stock_screener_heatmap.png): Color-coded grid view showing all stocks and their scores at a glance

Output Files

After running the screener, you'll find these files in the project directory:

stock_screener_results_20260107_123456.csv      # Results data (timestamp varies)
stock_screener_bar_chart.png                     # Bar chart visualization
stock_screener_heatmap.png                       # Heatmap visualization

Customization

You can easily customize the screener by editing stock_screener.py:

# Change the date range
START_DATE = '2025-01-01'
END_DATE = '2026-01-07'

# Modify the stock list
TICKERS = {"AAPL", "MSFT", "GOOGL", ...}

# Adjust scoring weights in the scoring functions
# For example, in ma50_score():
score = 0.5 * dist_score + 0.3 * slope_score + 0.2 * regime

Troubleshooting

Q: I get a TA-Lib installation error

• A: TA-Lib requires a compiled C library. See TA-Lib GitHub for platform-specific installation instructions.

Q: yfinance download is slow or fails

• A: This can happen during peak hours. The script includes multi-threading, but you can increase retry attempts by modifying the download_data() function.

Q: I get "module not found" errors

• A: Make sure you activated your virtual environment and ran pip install -r requirements.txt.

Q: Charts don't display

• A: If running on a headless server, comment out plt.show() lines in the script. Charts are still saved as PNG files.

Project Files

Stock Market Screener/
├── Stock Market Screener.ipynb    # Main interactive notebook
├── stock_screener.py              # Standalone Python script
├── requirements.txt               # Python dependencies
├── README.md                      # This file
├── .gitignore                     # Git ignore rules
└── stock_screener_results_*.csv   # Generated reports (after running)

Important Notes

• Score Range: All scores range from -1 (bearish) to +1 (bullish)
• Final Score Calculation: The final_score is the arithmetic mean of the three component scores
• Interpretation: Higher scores suggest better momentum and trend alignment
• Disclaimer: This is a technical analysis tool and should not be used as the sole basis for investment decisions
• Data Quality: Historical data is auto-adjusted for stock splits and dividends
• Real-Time Usage: This screener uses daily data; for intraday trading, modify the data frequency in stock_screener.py

Future Enhancements

• Additional indicators (Bollinger Bands, MACD, Stochastic Oscillator)
• Sector-based filtering and comparison
• Fully customizable date ranges via command-line arguments
• Integration with broader market indices
• Machine learning-based dynamic scoring weights
• Real-time alerts for significant score changes
• Web-based dashboard for interactive exploration
• Integration with brokerage APIs

Contributing

Found a bug or have an improvement? Feel free to fork and submit a pull request.

Author

Created as part of quantitative stock analysis research.

License

Personal use project. Feel free to modify and share for personal investment research.