Mai-Shan-Yun Vision ✨

Home page
Predictive model (forecasting)
Ingredient shipment
Pie charts for item category revenue per month
Most Popular items
Profit optimization by items

Inspiration

Our team wanted to help restaurants like Mai-Shan-Yun make smarter, data-driven decisions about their inventory and ingredient ordering. We have noticed that supply issues and ingredient shortages often lead to inefficiencies or lost sales. With the available data, we were inspired to develop a predictive tool that simplifies forecasting and operational planning.

What it does

Mai-Shan-Yun Vision is an all-in-one dashboard that forecasts ingredient demand, analyzes menu trends, and highlights potential inventory shortages before they happen. It combines forecasting models, visual analytics, and optimization tools to help restaurant managers:

Predict ingredient usage for upcoming months
Compare demand vs. shipment supply
Visualize top-performing menu items and revenue by category
Identify potential shortages with clear “action required” guidance

How we built it

We built the project using Python for data processing and forecasting, and Streamlit to create an interactive, user-friendly dashboard. Important technologies include:

Streamlit for the frontend interface and multi-page app structure
Pandas for cleaning and comparing ingredient and shipment data
Prophet for time-series demand forecasting
Altair and Matplotlib for clear and dynamic visualizations
Git/GitHub for version control and collaboration All datasets were processed locally and visualized directly in the Streamlit app to make insights easy to explore.

Challenges we ran into

We ran into a diverse array of challenges that helped develop our skills as programmers and data scientists:

Cleaning and merging multiple CSV datasets with inconsistent date formats and column naming
Ensuring the Prophet forecasts are aligned correctly with shipment schedules and constraints
Managing Streamlit’s layout and CSS customization for a clean and responsive design
Debugging module import paths when organizing the multi-page app structure

Accomplishments that we're proud of

The main accomplishment we are proud of is successfully building a multi-page Streamlit dashboard that integrates forecasting, insights, and optimization! Other great achievements we are proud of are:

Designed a smooth user interface with consistent styling, clear visuals, and a modern gradient theme
Implemented real predictive modeling (using Prophet) to forecast ingredient demand based on real sales data
Created actionable metrics that translate data into real restaurant insights

What we learned

How to integrate machine learning models (Prophet) into a Streamlit environment
Advanced data visualization with Altair and Pandas
The importance of structuring large Python projects for modularity and readability
Working together in a team, splitting tasks evenly, and coming together at the end to finish off the project strongly!