Project Overview
- Length: 1 month of preparation and 7 hours of Hackathon
- Content: MBA students teamed up with Data Scientists from GoodData. We were given public domestic flight data and 7 hours at the Hackathon to provide any insights that we’d like.
- Data set: 47 million rows, 4GB, CSV file
- Tools: We used R to build prediction model, Excel to deliver insights from historical data, and Google Slides for the presentation.
Our approach
- Team meeting – MBA students: Determined our analysis focus (flight delays? cancellation? others?) and developed hypothesis.
- Team meetings – MBA students and Data Scientists: Discussed about our analysis focus, agreed on the tools and amount of data we would use during the Hackathon
- Hackathon:
- I presented my idea of building a prediction model of flight cancellation.
- My team agreed, and GoodData members helped to clean the data and build a multiple regression prediction model.
- I came up with the creative recommendations about how flights search engines like Skyscanner could monetize from this model.
- We presented our insights in terms of “Descriptive”, “Predictive” and “Prescriptive” Analysis.
