MSY Challenge

Inspiration

We wanted to help Mai Shan Yun use real restaurant data to make smarter inventory and sales decisions. The idea was to bring data science and forecasting directly into everyday restaurant management.

What it does

The dashboard visualizes revenue, ingredient usage, and shipment trends while forecasting future demand. It helps minimize waste, prevent shortages, and plan restocks ahead of time.

How we built it

We built it using Python Dash, Plotly, Pandas, and Scikit-learn, integrating six months of restaurant data. The app was deployed on Render, with predictive models powered by Holt-Winters and Linear Regression.

Challenges we ran into

Limited data (only six months) made seasonality modeling difficult, requiring creative trend-based forecasting. We also had to ensure consistency across multiple Excel sources and maintain clean, real-time visual updates.

Accomplishments that we're proud of

We built a fully interactive and accurate forecasting system from real operational data. The dashboard clearly highlights trends that can directly improve restaurant efficiency and planning.

What we learned

We learned how to handle real-world messy datasets and translate them into business insights. We also gained experience deploying scalable, data-driven apps with clean visualization and forecasting logic.

What's next for MSY Challenge

Next, we plan to extend forecasting to item-level predictions and integrate live POS data. We’ll also enhance the interface for real-time alerts and mobile-friendly analytics.

Built With

• dash
• pandas
• plotly
• python
• scikit-learn
• statsmodels