Committing an experiment

Requirements

To commit an experiment, make sure the experiment file exists, under the following path in the data branch:
datasets/raw/<the_site_name>/<dataset_date>.csv

Example: datasets/raw/strasbourg/2016-12-28_18.52.44.csv

Extracting dataset informations

Still under the data branch, go the datasets/scripts/ directory and run the following:
python dataset_info.py <site_name> <dataset_date>

This will generate a json file under the following path:
datasets/processed/<site_name>/<dataset_date>/info.json

Publish the dataset

To publish the dataset, commit both the raw dataset and its processed info file. If the dataset if too large (>75Mb), you will no be able to push it on github. We are working on using GitHub (LFS).