The idea for this project came from observing how restaurants and food service operations often struggle to maintain the right stock levels. Ingredients can either run out, causing disruptions, or get overstocked, leading to waste. I wanted to create a tool that could intelligently forecast ingredient usage and recommend reorder points, using historical consumption data. The goal was to combine data science, time series analysis, and interactive visualization to help streamline inventory management.
Throughout this project, I gained experience in:
Time series forecasting using models like SARIMAX and Linear Regression (OLS).
Data preprocessing, including resampling monthly data and handling missing values.
Interactive dashboards with Streamlit and Plotly for dynamic visualization.
Inventory projection and safety stock calculation, connecting predictions to actionable insights.
I also learned the importance of designing code to handle edge cases gracefully: ingredients with no history, only one month of data, or outlier spikes in usage.
Log in or sign up for Devpost to join the conversation.