https://github.com/ericcccsliu/imc-prosperity-2?tab=readme-ov-file#round-3%EF%B8%8F%E2%83%A3
stanford 2023:
https://github.com/ShubhamAnandJain/IMC-Prosperity-2023-Stanford-Cardinal
The dataset comes from the IMC Prosperity trading competition, an algorithmic trading challenge. Data is stored locally in the imc-data/ directory, organized by round (round1 through round5)
The visualizations address trading strategy questions:
- Co-movement & volatility: How do assets move relative to each other? Which instruments are more volatile?
- Mispricing & mean-reversion: Is GIFT_BASKET fairly priced relative to its synthetic value (4×CHOCOLATE + 6×STRAWBERRIES + 1×ROSE)? When does the spread z-score revert to mean?
- Order-flow pressure: Does microprice (volume-weighted bid/ask) predict future mid-price movements?
- Volume imbalance: Does top-of-book volume imbalance correlate with short-horizon returns?
- Market quality: What are typical spreads and liquidity depths, and do they widen during mispricings?
Primary: Quantitative researchers, algorithmic trading engineers, trading competition participants, and students.
Secondary: Data scientists familiar with market data structures (order books, mid-prices, spreads).
This repository contains working Jupyter notebooks and Python scripts organized by round:
- round4-vis/round4_data_visualization.ipynb: Main Jupyter notebook analyzing Round 4 data with comprehensive visualizations
- round3-vis/spread_zscore_clean.py: Python script visualizing mean reversion strategy for gift basket spread trading
- round1-vis/mean_reversion_strategy_visualization.py: Python script for Round 1 strategy visualization
- round1-vis/visualize_mid_price.py: Python script for mid-price visualization
- Prerequisites: Python 3.x with pandas, numpy, matplotlib
- Install dependencies:
pip install pandas numpy matplotlib
Data files are expected in imc-data/round*/data/ directories relative to each script's location.