Flights Delays - demo

This repo contains notebook and scripts used to build a model to predict wether a flight will be delayed by 15min+ or not given the origin, destination, airline, date and time. Click here for a demo video

1. Objectives:

1. Build a model to predict how long a flight is to be delayed given flight date, origin and destionation airports as well as the airline
2. Build a user friendly web application

2. Data:

The following data are first collected in database

• [Flight delays data from transtats.bts.gov] (http://www.transtats.bts.gov/DL_SelectFields.asp?Table_ID=236&DB_Short_Name=On-Time)

• [Airports statistics from faa.gov] (http://www.faa.gov/airports/planning_capacity/passenger_allcargo_stats/passenger/previous_years/)

3. Tools:

• Python
• SQL
• HTML, CSS, Javascript, D3JS

4. Strategy:

The dataset contains about 30 features about a flight and the airports. However, at the time of booking, the only information available to a user (passenger) are date, origin, destination and airline. In order to take advantage all of the information provided by our data sources, I first analyze all features about airports separately from features about flights and then generate clusters of similar airports snd flights.

1. [Airports analysis] (https://github.com/kthouz/Flights_Delays/blob/master/01_1_Airports_Analysis.ipynb): run statistical analysis and cluster similar aiports together

2. [Flights analysis] (https://github.com/kthouz/Flights_Delays/blob/master/01_2_Flights_Analysis.ipynb): run statistics and cluster similar flights together

3. Delays analysis: build predictive model

5. Application:

Build a front-end of a web application using HTML, CSS, JavaScript and D3JS