Team Airtraffic is creating insightful visualizations of air traffic from / to Zurich Airport and predicting delays
So far, we have:
- conducted exploratory data analysis and produced many descriptive visualizations to get to know the data and get a first understanding of which aspects of it might be relevant and interesting
- experimented with a broad variety of modelling approaches (linear approaches, binary logistic regressions, ridge regression...) to predict delays
Next steps :
- Try further prediction-approaches (baynesian, random forest e.g.)
- Create some further insightful visualizations
- Discuss what the model could be used for (flight delay prediction app? )
Flight data for ML-prediction challenge at the TWIST2018-Hackdays
This repository contains data of planed and effective flight arrival and departure-times from / to the airport of Zurich for the entire year 2017.
The R-Script data_enrichment.R contains the code used to add the airport coordinates, calculate approximative flight-distances, time-differences between scheduled and effective flight-times as well as information on weather-conditions around the airport provided by meteoswiss. The dataset does not yet contain information on general weather and atmospheric conditions other than those at the local scale.
The RDS-file twist_zrh.RDS contains the resulting R-dataframe. For those relying on other tools than r, a csv-version of the dataset is also available: twist_zrh.csv
metadata.txt contains a detailed description of the variables contained in the file.
Further sources for weather data at global scale or atmospheric conditions:
meteorogical conditions on the ground via NASApower: https://adamhsparks.github.io/nasapower/
atmospheric conditions via: http://www.wmo.int/pages/prog/www/GOS/ABO/data/ABO_Data_Access.html#gts