Put CY2014.csv and drug_table_jun2015b.csv (or symlinks) into ./original-data.
Do the following once:
cat ./original-data/sample-raw-data.csv | cut -d, -f 1,8,10,12,13 > ./original-data/minimal-dataset.csv
Parse file and grab drug codes of interest into a pickle object:
python build-drug-code-set.py
create list of at-risk patients, grab a collection of all provider ids to be used later for subsetting, and bundle our data of interest into a pickle of PatientRecord objects
pass a --max_records arg to limit or for testing
python process-data.py --max_records=100000
select a random subset of providers:
python subset-providers.py --providers=1000
build the network:
python build-network.py
calculate the stats:
python calculate-stats.py