This assignment will have you do a little machine learning. Giving historical airplane on time performance data, you will offer suggestions for two-hop flights that minimize the chance of missing a connection.
Fine print:
- Input: A list of tuples of Year, Month, Day, Origin, Destination (e.g. 2001, 09, 11, DEN, DCA) where Year is the same for all tuples
- Ouput: a list of proposed flights (e.g. [(2001,09,11, 1256, 1222, UA, DEN, ORD),(2001,09,11,1910,UA, ORD, DCA)])
- Scoring: 1 point per flight if the flight arrives on time at the destination, -100 per proposed flight that does not arrive on time or is cancelled
- You can use data form previous years to build a model of flight delays, you are not allowed to use data from the Year requested in the model
- Do not return direct flights
- Connections must have at least 45 minutes between landing and takeoff, and no more than 12 hours
- The PDF report file should include details about the implementation as well as a study of the performance and accuracy of the solution. A detailed breakdown of what each team member worked should be added. Grading will include the quality of the report. All projects will be code walked.