Abstract - This paper addresses the challenge of accurate demand forecasting and inventory management in the food delivery industry. Sales forecasting is crucial for the success of independent restaurants and restaurant chains. To mitigate the risks of unpredictable consumer demand and perishable raw materials, we have developed a website that utilizes an appropriate machine learning model to forecast daily and weekly demand for raw materials. Our model is trained on a dataset that includes information about fulfilment centres and meal categories. We have utilized the Gradient Boost Regression model for food demand forecasting, achieving a training accuracy of 92% and a testing accuracy of 83%. Our web application also includes a separate inventory management section for restaurant owners to manage their available inventory, customers, and orders. This project provides a generalized model that can be adapted by other companies or services to their data. The project utilizes a tech stack of HTML, CSS, Javascript, Bootstrap, Python framework Django, SQLite, and the Gradient Boost Regression model. Overall, our project offers a valuable solution for restaurants and food delivery companies to better manage their inventory and satisfy customer demand.

 Keywords - Machine Learning approach, Food demand forecasting (10 Times new Roman)

| DOI: 10.17148/IARJSET.2023.10568