Project Description
Description: The purpose of our project is to analyze different features that may affect the departure of flights and predict the fight delays. We choose two airports that are close to us: BWI and DCA, and hope to help people to form a reasonable expectation of possible delays in their next trip.
Data Source: We get our data from the US Department of Transportation’s Bureau of Transportation Statistics website. We select 1-year data of flights departing from BWI or DCA in 2017. All variables we think may affect the departure of flights are downloaded first and then processed differently based on their properties.
Summary:
- We analyze the effects of different features on flights’ departure on-time status for BWI and DCA airports.
- Several types of classifiers are trained before and after PCA transformation to original datasets.
- Among these classifiers, Decision Tree provides relatively better prediction results for both BWI and DCA airports.
- Given conditions, like weather, destination city and airline, we could predict flights’ on-time status with over 82% accuracy, which may help passengers form reasonable expectation of their flights’ departure time.