Case Study: Optimising Operations and Deliveries
This project presents a data-driven case study on improving coffee shop operations using historical order data and event logs.
The analysis identifies process bottlenecks, forecasts customer demand, and provides recommendations for optimising deliveries, staffing, and customer experience.
- Process Mining: Analysed event logs to identify bottlenecks in the order-to-delivery process.
- Forecasting: Predicted demand patterns based on 12 weeks of historical data.
- Prescriptive Analytics: Recommended strategies for staffing, courier allocation, and customer engagement.
-
Process Bottlenecks
- The Call Customer step created loops in 24% of cases, increasing delays.
- The Coffee Reaches → Payment step averaged 20 minutes delay in 71% of cases.
-
Order Trends
- Demand consistently peaked on weekends.
- Highest customer demand occurred between 8–10 AM and 1–3 PM.
- Forecasting model deviation: ±17 orders/day, indicating inventory risks.
-
Operational Constraints
- Courier cap of 10 per shift limited fulfillment on long-distance deliveries.
- High-volume days (e.g., 518 orders on 9th June) exceeded courier capacity.
- Increase courier capacity during high-demand hours and weekends.
- Implement automated route planning with real-time traffic data.
- Introduce loyalty programs to drive weekday and off-peak orders.
- Offer wider digital payment options to reduce checkout delays.
- Automate order handling to eliminate manual customer calls.
- Apply peak-hour surcharges to offset courier costs during demand spikes.
- Excel: Initial data exploration, cleaning, and summary analysis.
- Celonis (Process Mining): Identified process bottlenecks and inefficiencies in the order-to-delivery workflow.
- Python: Data analysis and forecasting (pandas, numpy, matplotlib, seaborn)
- Power BI: Built interactive visuals to communicate demand trends and support decision-making.