SmartStock
- #Marketplace
Automated Inventory Forecasting System.
- Machine Learning
Impact
- The main goal is to simplify inventory forecasting, reduce manual data entry, and use the automated solution for effective inventory management.
- Eventually, the solution will be integrated into the client’s services and can be monetized as part of their offerings.
Services we provided
Forecasting solution for efficient inventory management
Tech Stack
Pytorch
Python
Streamlit
Pandas
Scikit-Learn
FastAPI
GitHub
Challenges and Solutions
🧐 Challenges
- Сreate a working ML model integrating into the existing platform.
- The system should be capable of handling a diverse range of SKUs with varying selling rates.
- To ensure forecast accuracy and reliability.
- To design a scalable solution for future growth.
💡 Solutions
- Developed an automated system that uses data from a third-party company. This data includes info about product availability and sales rates in their store.
- The system predicts future orders for each store by analyzing the current store inventory, inventory in SmartStock, and past sales history.
User flow
1. User uploads an Excel file with data.
2. System generates an Excel file with processed and forecasted columns: “Est days remaining”, ‘Need’ (calculated based on the current inventory and prior sales history); ‘Landed’, etc.
3. It enables store-specific future order forecasting.