Scripts used to generate the analysis and figures for the paper and the SI.
data contains all original or intermediate data files needed for the analysis. These files can either be downloaded by running the download scripts or by running the python scripts.
download_submissions.sh downloads all unique final submissions from the top 400 teams. These are needed to be able to run the preprocess.py script.
preprocess.py parses the submissions downloaded by download_submissions.sh and creates a pickle with the processed data. This pickle can be downloaded with the download_preprocessed_data.sh script.
download_preprocessed_data.sh downloads the output from the preprocess.py script.
output will contain the output from the blend.py and make_submission_plots.py scripts.
blend.py fits the weights of two different ensembles, writes auxiliary data files to the data folder and writes the scores and weights to the output folder.
make_submission_plots.py creates and writes the plots related to the submitted predictions of participants shown in the paper and supporting information to the output folder.
make_auxiliary_plots.py creates and writes the remaining plots shown in the paper and supporting information to the output folder.
download_duplicate_submissions.sh downloads submissions that were a duplicate of another submission and was therefor removed from analysis.