ampli/py/agg.py

# Aggregrate given clusters
import pandas as p

def aggregator(widedf, clusdf):
    clusters = clusdf['cluster'].unique()
    clusters.sort()
    aggv = {}
    qlow  = lambda x: x.quantile(0.250)
    qhigh = lambda x: x.quantile(0.750)
    for c in clusters:
        icps = clusdf[clusdf.cluster == c].icp_id.values
        subdf = widedf[icps]
        agged = subdf.agg(func = 'median', axis = 1)
        print(agged)


def main():
    wd = p.read_pickle("../data/2017-5k-wide.pkl")
    cd = p.read_pickle("../data/5kclustable.pkl")
    aggregator(wd, cd)



if __name__ == "__main__":
    main()