Implemented and executed the Round 5 EDA notebook.

Created:

• round5_relative_value_eda.ipynb
• round5_relative_value_eda.py

Generated:

• outputs/: 11 files, including cleaned data, features, summaries, residuals, diagnostics, correlations, and research_summary.txt
• figures/: 120 files, including 10 Plotly HTML dashboards

Verification:

• Loaded all 50 products across days 2, 3, and 4.
• Cleaned trades exclude buyer, seller, trader_id, and identity-like fields.
• Executed the notebook successfully with jupyter nbconvert.
• python -m py_compile round5_relative_value_eda.py passed.

Top summary from the run:

• Strongest common level movement: SLEEP_POD
• Highest average return correlation: OXYGEN_SHAKE
• Cleanest residual groups: SLEEP_POD, SNACKPACK, ROBOT
• Most negative h=10 residual gamma examples: SNACKPACK_RASPBERRY, PEBBLES_S, MICROCHIP_OVAL
• Widest relative spreads are mostly SNACKPACK products