import pandas as pd
import numpy as np

pd.options.display.max_columns = 5200
pd.options.display.max_rows    = 5200

pj = '/Users/danielmsheehan/Desktop/data/'
pj = '/Users/danielmsheehan/Dropbox/data/'
pj = '/Volumes/Hotel/Dropbox/data/'

df = pd.read_csv(pj+'processing/buildings/cb2010_building_0913_int.csv', dtype={'geoid':object}).rename(columns=lambda x: x.lower())

df

dfg = df.groupby('geoid')

dfg.head(20)

dfg_desc = dfg['bin'].describe()

dfg_desc.head(20)

dfg_desc.to_csv(pj+'processing/buildings/cb2010_building_0913_int_describe.csv')

df = pd.read_csv(pj+'processing/buildings/cb2010_building_0913_int_describe.csv', dtype={'geoid':object}, header=None)

df.columns = ['geoid', 'stat','value']
df.head(50)

df_cnt = df[(df.stat == 'count')]

df_cnt.columns = ['geoid', 'stat','count']

df_cnt['countbldg'] = df_cnt['count'].astype(int)
df_cnt = df_cnt[['geoid','countbldg']]

df_cnt.to_csv(pj+'processing/buildings/cb2010_building_0913_int_describe_cnt.csv', index=False)

df_cnt.head(50)