Efficient Food Delivery Management: Predicting Delivery Times for Improved Customer Experience

With more and more customers depending on delivery services for their meals, the food delivery business has grown significantly in recent years. One critical aspect that directly impacts customer satisfaction is the accurate prediction of delivery times. Inaccurate delivery time estimations can lead to dissatisfaction and negatively impact the overall customer experience. This paper addresses the challenge of efficient food delivery management by proposing a predictive model for estimating delivery times. Through meticulous data preprocessing and feature engineering, the paper enhances the dataset, ensuring its suitability for machine learning model development. Exploratory Data Analysis (EDA) aids in the selection of pertinent predictors by offering insightful information about the connections between various parameters. Many machine learning methods, such as Linear Regression, Lasso, Decision Tree, XGBoost, and RandomForest, are applied during the predictive modeling phase. GridSearchCV hyperparameter adjustment is used to maximize the performance of every model. The results showcase the strengths and weaknesses of each model, with a focus on the selected best-performing model. The discussion section interprets the findings, comparing the models' predictive capabilities and their potential implications for food delivery management. Recommendations are provided for implementing the predictive model in real-world scenarios, along with suggestions for future research directions. The study concludes with a summary of key insights and underscores the crucial role of accurate delivery time predictions in enhancing customer satisfaction and overall service efficiency.