flight_arrival_delay_predict

Predicting flight arrival times based on regression models This report presents the process of analyzing and building predictive models for flight delays in the United States, based on the On-Time Reporting Carrier On-Time Performance dataset for 2022. The team conducted data collection, preprocessing, exploratory analysis, and visualization on over 3.9 million records, focusing on factors affecting delays such as airline, delay causes, time, and flight routes. The analysis results show that most delay time originates from airline-related causes and late-arriving aircraft, with carriers like JetBlue, Frontier, and Allegiant exhibiting the highest delay rates. The team implemented simple linear regression, multiple linear regression, and polynomial regression models to predict arrival delay based on features such as departure delay and departure time, with the simple linear regression model achieving the best performance (R2 score of 0.93673). Additionally, an interactive dashboard was developed using Streamlit to visualize data and provide quick delay predictions for individual flights. The findings offer an overview of the current state of flight delays in the US and support airlines and passengers in making more informed decisions when choosing and managing flights.