Saucy Sales

Project Logo

About the Project

Saucy Sales was born from the realization that many pizza shops struggle with overproduction and food waste while missing opportunities to optimize inventory and staffing. Inspired by industry reports and real challenges—like $1.3B in wasted food annually and inaccurate demand forecasts—this project was designed to provide actionable insights and help pizza operators turn data into better decision making.

What it does

Forecasting: Uses historical sales data and AI-driven models to predict future sales trends.
Interactive Dashboards: Visualizes trends via responsive charts (using Chart.js and Next.js), helping users understand peak hours, seasonal trends, and menu performance.
Menu Simulation: Allows operators to test changes on menu items and see their potential impact on sales and inventory management.

How We Built It

Frontend: Developed a modern, responsive dashboard using Next.js 14, React, and TypeScript. Interactive components display charts for net gains, seasonal trends, peak order hours, and top menu items.
Visualization: Leveraged Chart.js for dynamic and interactive data visualizations that make insights easy to understand.
Backend & Data: Integrated data from CSV files and used Python (with libraries like Prophet, pandas, and CatBoost) for robust forecasting. Cleaned and merged multiple datasets to ensure accurate predictions.
APIs & Integration: Incorporated REST API endpoints (via Axios) to communicate between the forecasting models and the frontend interface.

Challenges We Ran Into

Data Quality: Cleaning and merging datasets from various sources posed significant challenges.
Forecast Accuracy: Tuning the AI models to account for seasonal dips and spikes required careful experimentation and validation.
Responsive Design: Ensuring that interactive charts render correctly on all devices and maintain performance under load.

Accomplishments That We're Proud Of

Achieved a forecasting accuracy that provides actionable insights for inventory and staffing.
Reduced predicted food waste by optimizing menu offerings and stock levels.
Built an engaging and intuitive interface that empowers pizza operators to make data-driven decisions.

What We Learned

The importance of robust data cleaning techniques before building forecasting models.
Merging front-end interactivity with back-end AI and data science can unlock powerful insights.
Responsive design and user-centric features are critical for dashboard success in real-world environments.

What's Next for Saucy Sales

Expand the model to support multiple locations and more extensive datasets.
Integrate real-time data updates and alerts for inventory and staffing needs.
Enhance the AI model with additional factors (e.g., weather data, local events) for even better predictions.

Presentation