lens_example.py

# from http://docs.sherlockml.com/libraries/lens/getting_started.html
import lens
import pandas as pd

df=pd.read_csv('300_casas_renam.csv.gz', compression='gzip')
df=df[['id_x','text_x','hext_x','time','tint_y','hint_y','Co2_y','Ruido_y']]
df=df.rename(columns={'id_x':'casa','text_x':'TempExt','hext_x':'HumedadExt',
            'hint_y':'HumedadInt','Co2_y':'CO2',
            'Ruido_y':'Ruido','tint_y':'TempInt'})

df['Humedad']=pd.cut(df['HumedadInt'],5, 
                labels=['seco','bajo','medio','alto','humedo']).astype(str)

ls = lens.summarise(df, scheduler='sync')   # 1m19s for 100K casas on the basic sherlock instance
ls.to_json('lens_summary.json')
explorer = lens.explore(ls)
explorer.describe()
explorer.correlation_plot()