The warehouse team at the Houston Food Bank needed to prevent inventory shortages and overages so that they are able to fill all scheduled orders on-time and minimize food waste. Data for the amount of inbound inventory, outbound inventory, and inventory on hand was disparate and existed in different reports across the organization.
Inventory data was also being collected and recorded at different levels of granularity across different systems. Because there was no shared level of data to aggregate the different systems, the Houston Food Bank had not been able to project inventory levels in an actionable way.
Our team used natural language processing and machine learning to label uncategorized data and unify inventory data across systems. Our model was trained on historical warehouse data and was able to accurately label inventory 97% of the time.
By labeling uncategorized data and unifying different systems, the project allowed the Houston Food Bank to aggregate the data at a shared level of inventory and predict when specific food categories might experience shortages or overages. This allowed Houston Food Bank staff to proactively adjust their operations to prevent these shortages and overages.
The Houston Food Bank is the largest food bank in the country. Last fiscal year the Houston Food Bank provided 159,000,000 meals to families in need in the 18 southeast Texas counties they serve. On a weekly basis, they serve over 92,000 families. Efficient management and distribution of their warehouse inventory is paramount to ensuring that they are able to serve the 1.1 million food insecure people in southeast Texas effectively. Our dashboard empowers the Houston Food Bank to be able to predict shortages and overages for the first time.