Overview
Small businesses lose money when they misjudge how busy a day will be. ServeSmart predicts daily customer counts so owners can staff and stock accordingly and reduce waste.
How it works
- A recurrent neural network built in TensorFlow predicts the next day's customer count
- Predictions use factors like the previous day's customers, the day of the week and the weather
- A React front end and a Flask back end serve the predictions
Result
Our team won 2nd place in the small business division at Hack MHS, out of 90 participants overall.