The goal is to predict arrival delays of commercial flights. Often, there isn't much airlines can do to avoid the delays, therefore, they play an important role in both profits and loss of the airlines. It is critical for airlines to estimate flight delays as accurate as possible because the results can be applied to both, improvements in customer satisfaction and income of airline agencies.
We will be working with data from air travel industry. We will have four separate tables:
- flights: The departure and arrival information about flights in US in years 2018 and 2019.
- fuel_comsumption: The fuel comsumption of different airlines from years 2015-2019 aggregated per month.
- passengers: The passenger totals on different routes from years 2015-2019 aggregated per month.
- flights_test: The departure and arrival information about flights in US in January 2020. This table will be used for evaluation. For submission, we are required to predict delays on flights from first 7 days of 2020 (1st of January - 7th of January). We can find sample submission in file sample_submission.csv
The data are stored in the Postgres database. You can see the information about the hostname and credentials in Compass.