The Problem
Products that expire before being consumed must be destroyed Significant source of waste
Our Solutions
- Leveraging order data to suggest optimal purchase patterns
- Leveraging order data to predict the orders a wholesaler will make in the next month
- Leveraging order data to predict margin of time before expiry on orders
Optimal Purchase Patterns
- How to improve efficiency of orders from wholesalers?
- Many wholesalers order the same item multiple times per week — identify and combine into larger orders.
- Can set desired order frequency (ie. monthly) and prioritize wholesalers and products to focus on
Combining Shipements
- Across different products, which shipments could be combined?
- We developed a method to identify clusters of products where purchases are correlated but shipped separately
- For each wholesaler, we Identify opportunities for combining orders and reducing the number of total shipments
Models
- Random Forest
- Linear (Lasso) Regression
Results
- Average r-squared score for linear regression 0.87/1.00
- Some features are more important than others
Conclusions
Conclusions from expiry margin prediction problem:
- Develop drugs with longer expiry dates!
- Some (anonymized) molecules and products seem to have problems in the supply chain (Ask us about which ones!)
Log in or sign up for Devpost to join the conversation.