Data Analyst · intermediate
Supply Chain Optimization & Demand Forecasting Tool
A predictive analytics tool that forecasts demand and identifies supply chain bottlenecks via an interactive BI report.
Project overview
Supply chain resilience is a massive corporate priority right now; this project proves you can solve high-stakes operational problems and save companies significant costs.
Real-world use case
Logistics and manufacturing firms (like FedEx or Tesla) managing global inventory and trying to minimize stock-outs and overstocking.
Tools and technologies
- Python
- Scikit-learn
- Prophet
- PostgreSQL
- Power BI
Skills you can demonstrate
Time-Series Forecasting, Predictive Modeling, Supply Chain Analytics, SQL
Build roadmap
- Get historical sales data
Use a Kaggle retail dataset or generate three years of daily sales with weekly and holiday seasonality. Load it into PostgreSQL with a proper date dimension. - Explore seasonality patterns
Plot daily, weekly, and yearly views with pandas and Plotly. Mark holidays and anomalies — a forecast is only as good as what you understood about the past. - Build the Prophet model
Train Prophet with holiday effects and produce 90-day forecasts with confidence intervals. Hold out the last 90 days so you can measure accuracy instead of guessing. - Compare to other models
Train ARIMA and a simple LSTM on the same split, score all three with MAPE, and write one paragraph on which won and why. That paragraph is the portfolio. - Build the stock-out risk query
SQL that joins inventory, lead times, and forecasted demand to flag SKUs likely to run out inside their lead time. Return a ranked list, not a dump. - Build the Power BI report
Forecast visuals, a risk heatmap by SKU and region, and a what-if panel for lead time. Present it as the page a supply chain manager opens every Monday.
Ways to extend the project
Use an AI agent to simulate "What-If" scenarios — like the impact on inventory if a major shipping port closes or a supplier is delayed.
How to showcase your work
- GitHub: SQL scripts + Python notebooks, fully documented.
- LinkedIn: Carousel of insights and cost-saving metrics.
- Resume: Focus on "cost-saving" and "efficiency" metrics.