Hamza Salah
Data Science (M.S. graduate), passionate about building ML and forecasting solutions across diverse domains.
Operational Schedule & Capacity Analysis
(NBA Dataset Case Study)Project Overview
Using NBA schedules (2014–2025) as a proxy for high-frequency operational environments, this project models how compressed timelines impact performance. By identifying “dense stretches” (such as back-to-backs or 4-events-in-6-days), the system quantifies the degradation in efficiency and win-rates.
The analysis uses regression modeling to offer strategies for optimizing workload, managing fatigue, and improving resource allocation during peak operational windows.
Interactive Schedule Density Visualizer
Navigation: Drag to zoom, shift+drag to pan, double-click to reset, hover for details.
Key Operational Insights
- Performance Degradation: Dense scheduling clusters (e.g., back-to-backs) correlate with measurable declines in output quality (specifically defensive efficiency), serving as a proxy for workforce fatigue.
- Trend Analysis: The frequency and severity of these “high-stress” operational windows have trended upward over the last decade (2014–2025).
- Resource Allocation: Analysis suggests that “deeper rotations” (analogous to flexible staffing models) significantly mitigate the negative impacts of dense scheduling.
Code and Data Integration
- Datasets: - NBA schedule data (2014–15 to 2024–25):
- Data Cleaning & ETL Pipeline: - Standardization: Normalized date formats and sorted chronological workflows.
- Filtering: Segmented schedules by organizational unit (Team) and fiscal year (Season).
- Feature Engineering: Calculated logic for “Constraint Stretches” (Back-to-Backs, 4-in-6s).
- Normalization: Adjusted stretch counts to an 82-game baseline to ensure fair year-over-year comparisons.
Technologies Used
- Analysis & Modeling: Python (Pandas, NumPy, Scikit-learn, Statsmodels)
- Visualization: Plotly (Interactive Dashboards), Matplotlib
- Environment: Jupyter Notebook for data analysis and visualization