Statistical Approach to Candlestick Patterns

Overview

This repository offers a data-driven toolkit for analyzing candlestick patterns in stock market data. It provides robust methods for detecting patterns, visualizing them across long-term data, and statistically analysing their effectiveness.

For a detailed walkthrough of the analysis and methodology, check out our Medium articles:

1. Candlestick Patterns in Trading: A Data-Driven Journey (Part 1)
2. Candlestick Patterns in Trading: A Data-Driven Journey (Part 2)

Features

• Programmatic detection of various candlestick patterns
• Interactive visualization of patterns on stock charts
• Statistical analysis of pattern returns
• Monte Carlo simulations for assessing pattern effectiveness
• Pattern effectiveness analysis

Getting Started

1. Clone the repository:

   git clone https://github.com/CharlieNestor/statistical_analysis_candlestick_patterns.git
   

2. Install the required packages in a new virtual environment:

   python -m venv venv
   source venv/bin/activate
   pip install -r requirements.txt
   

3. Run the Streamlit app:

   streamlit run pattern_streamlit.py

Usage

The Streamlit app provides an interactive interface for analysing candlestick patterns. It is recommended to set the background color to white for better visibility.

1. Select a stock from the dropdown or enter a custom ticker.
2. Choose a candlestick pattern to analyze.
3. Use the checkboxes to display:
   • Stock chart with pattern occurrences
   • Examples of individual pattern instances
4. Click "Run Simulation" to perform statistical analysis.
5. View the results, including:
   • Pattern vs. Base Case Performance chart
   • Significance Heatmap

File Structure

• patterns.py: Contains functions for programmatically define various candlestick patterns
• loader.py: Utilities for loading and preprocessing stock data
• plots.py: Functions for creating interactive visualizations
• analysis.py: Functions for statistical analysis of pattern returns
• pattern_stock.py: Defines the PatternStock class which will be used int the Streamlit app
• pattern_streamlit.py: Streamlit app for interactive pattern analysis
• Pattern_Medium_01.ipynb: Jupyter notebook showcasing the code journey related to the first Medium publication
• Pattern_Medium_02.ipynb: Jupyter notebook showcasing the code and analysis journey described in the second Medium publication
• Pattern_playground.ipynb: Jupyter notebook for those who wants to directly play with the patterns and the functionalities

Future Work

This repository is part of an ongoing project. Future updates will include:

• Expansion of the pattern library
• Integration of machine learning techniques for pattern recognition
• Development of a backtesting framework for trading strategies based on candlestick patterns
• Optimization of the Streamlit app for better performance and user experience

I'm committed to evolving this project into a comprehensive, easy-to-use solution for candlestick pattern analysis in stock trading.