Abstract / Introduction

Accurately predicting customer demand is one of the biggest challenges faced by restaurants and food outlets. Underestimating demand can lead to shortages and dissatisfied customers, while overestimating often results in unnecessary food wastage. At present, many small and medium-sized restaurants still rely on manual estimates and guesswork to manage their food supply chain, which frequently causes inefficiencies.

Efficient food supply chain management is essential for minimizing waste, ensuring timely availability of items, and enhancing customer satisfaction. The goal of this project is to design and implement an AI-enabled system that can predict daily and weekly customer counts and their food demand using data-driven approaches. By providing accurate demand forecasts, this system will support better inventory planning, reduce wastage, and optimize overall supply chain operations.

Functional Requirements:

Data Collection

• The system shall collect historical data of customers (daily and weekly counts).
• The system shall record food items sold, quantity, and time of purchase.
• The system shall allow manual data entry for smaller outlets without POS systems.

Data Pre-processing

• The system shall clean raw data by removing duplicates, handling missing values, and normalizing entries.
• The system shall support data transformation (e.g., grouping by day, week, or food category) among others.
• The system shall allow users to upload CSV/Excel files.

Prediction

• The system shall predict the number of customers for the next day and the upcoming week.
• The system shall forecast the demand for specific food items based on historical consumption.
• The system shall allow the user to choose between multiple prediction models (e.g., Linear Regression, ARIMA).

Visualization & Reporting

• The system shall generate graphical charts (line, bar, pie) showing past trends and predicted demand.
• The system shall provide downloadable reports in PDF/Excel format.
• The system shall display daily and weekly prediction dashboards for quick insights.

Inventory Planning Support

• The system shall suggest approximate quantities of raw materials required based on predicted demand.
• The system shall generate alerts if the predicted demand is significantly higher or lower than usual.

User Management

• The system shall allow different types of users (e.g., admin/manager, staff).
• The system shall support secure login and authentication.
• The system shall maintain a history of predictions for future reference.

System Maintenance & Updates

• The system shall allow retraining of the prediction model when new data is added.
• The system shall automatically update predictions when the dataset is updated.