To follow the exercise in class, you can install R https://cran.r-project.org/ (and RStudio https://www.rstudio.com/ recommended for Win/Mac) or try to use RinRuby from your Ruby console and download the following files from the folder
💾 train.csv and 💾 test.csv
Two files with 70% (train) and 30% (test) of the data on flight delays. The data looks like this:
| ARR_DEL15 | DAY_OF_WEEK | CARRIER | DEST | ORIGIN | DEP_TIME_BLK |
|---|
0| 2| AA| LAX | JFK | 0900-0959
0| 2| AA| LAX |JFK |1200-1259
0| 2| AA| SFO |JFK | 0700-0759
1| 2| AA| JFK |SFO | 1100-1159
0| 2| AA| SFO |JFK | 0800-0859
......
📄 predict_flightdelays.RMD is a R html markup file with the code to readin the data, train a model, test it on the test dataset and evaluate the statistics. If you don't have R, you can go to https://github.com/mbbrigitte/Ruby_Talk_Material/blob/master/predict_flightdelays.md to see the mark up online (or http://rpubs.com/mbbrigitte/ML_FlightDelayPrediction).
Data you don't need but are used in the talk
- Original_data.zip obtained from http://1.usa.gov/1KEd08B
- prepare_flightdata.RMD is the code to clean and tidy the data and split them into train and test, and save them as csv
- Cluster_presentation.RMD is the short code presented to perform hierarchical clustering on random data.
EOF