Muhammad T

Data Scientist

GitHub

Sep 2022 - Present • 4 yrs 1 mo

Data doesn't explain itself. That's where I come in. Over the past four years I've worked on data problems across supply chains, city traffic systems, sales operations, and customer behavior — not as classroom exercises, but as real end-to-end projects where the goal was always the same: find what's actually happening inside the data and say it clearly. In supply chain I worked with a large logistics dataset, engineered new features from scratch, built anomaly detection to flag unusual patterns, and traced where delays and losses were actually coming from — not just where people assumed they were. In traffic prediction I built a forecasting system tuned specifically for Saudi city conditions — Friday prayer time drops, Ramadan schedule shifts, sandstorm events. I compared multiple ML models, picked the best performer, and deployed the final system via FastAPI with live prediction, batch processing, and health check endpoints. It runs. In sales I analyzed chocolate sales data to surface which products, regions, and salespeople were actually driving results versus which ones just looked good on paper. I've also studied customer churn — how businesses identify customers likely to leave and what signals in the data predict that behavior before it happens. My tools: Python, Pandas, Matplotlib, Seaborn, Scikit-learn, FastAPI. I work with CSV and Excel files. You get charts, a written summary, and plain English — no jargon, no confusion. If your data has a story to tell, I'll find it.