Magic Manager

Inventory Dashboard UI

Inspiration

We were inspired by all the waste many restaurants experience due to poor inventory management and predictions.

What it does

This is an Ai powered restaurant inventory management dashboard. It uses machine learning models to predict user demand and helps managers make decisions on what items to discount and when to stock up on items to optimize inventory flow.

How we built it

We used Sklearn to create random forest models to help predict user demand. We created a FastAPI backend to calculate metrics to determine stockout rates as well as make APIs to connect a NextJS frontend to a SQLite database.

Challenges we ran into

We had problems with UI optimization and ML training. For UI optimization, the team discussed multiple designs and layouts to help optimize the user experience of the dashboard. We eventually came to a consensus and tried out various refinements to our original design. For the ML training, we had issues with data quantity. We had low data quantity and decided to use stronger ML models and generate some synthetic data for training purposes.

Accomplishments that we're proud of

We are proud of the ML models we built as well as the magic themed dashboard.

What we learned

We learned how to process data and use Sklearn to create various models. We also learned how to make a better frontend using TailwindCSS.

What's next for Magic Manager

We want to make more accurate ML models to better predict user demand. We would also like to work on the UI some more to make it more informative. We would further like to improve the speed and efficiency of the backend.