Flight Delay Predictor

Are full service airlines worth the extra cost?

For the final project in the Data Analytics and Visualization bootcamp, we worked in groups of 4–5 people to combine all of the tools and skills learned in the six-month course to analyze a dataset of our choice using machine learning, then visualize the results.

Our team chose to analyze flight delay data from the Hartsfield-Jackson Atlanta International Airport (ATL) using a logistic regression model to predict if flights would be delayed based on historical trends.

We chose ATL because it is the busiest airport in the world and would be able to provide large amounts of data while allowing us to limit the dataset’s variables to prevent large variables from impacting the results. Our dataset was further narrowed by only analyzing data from the first 8 months of 2022 and limiting our analysis to seven airlines:

  • “Full Service” airlines: American Airlines, Delta Airlines, and United Airlines
  • “Low Cost” airlines: JetBlue and Southwest
  • “Ultra-Low Cost” airlines: Frontier and Spirit

Ultimately, the “BIG QUESTION” we wanted to answer with our analysis is if the cost of an airline was directly related to its on-time probability.

0
Pieces of Data
0
Accuracy Score

Tools Used

Excel
POSTGRES
python
pandas
tableau

Analysis and Visualizations

Delays by destination airport

On-Time Probability by Carrier

Carrier
Delays
Total Flights
On-Time
American Airlines 572 3,577 84.01%
Delta Airlines 18,319 102,520 82.13%
United Airlines 389 1,717 77.34%
JetBlue 339 1,043 67.50%
Southwest Airlines 3,217 15,094 78.69%
Frontier Airlines 1,044 3,042 65.68%
Spirit Airlines 1,086 4,315 74.83%

On-Time Probability by Category

Full Service
81.16%
Low Cost
73.09%
Ultra-Low Cost
70.26%

Interactive Tableau Dashboard

Google Slides presentation

Related Projects
Clear Filters
Roll for Sandwich Data Analysis
Data Visualization: Unification of Great Britain
Data Visualization: Unification of Great Britain