Aggrigate impressions per day from exported twitter analytics csv with pandas

aggrigate.py

import pandas as pd

df = pd.read_csv('tweet_activity_metrics_example_20170314_20170412_ja.csv', header=None).loc[:, [3, 4]]
df = df.drop(0)
df[3] = pd.to_datetime(df[3])
df[4] = pd.to_numeric(df[4])
result = df.groupby([df[3].dt.year, df[3].dt.month, df[3].dt.day])[4].sum()

print(result)
# date           sum of impressions
# 2017  3  14    127.0
#          15     66.0
#          16     31.0
#          17    129.0
#          18    175.0
#          19    128.0
#          20    109.0
#          21    157.0
#          22    113.0
#          23    106.0
#          24    153.0
#          25    116.0
#          26    124.0
#          27    199.0
#          28    155.0
#          29    146.0
#          30    196.0
#          31    230.0
#       4  1     164.0
#          2     183.0
#          3     282.0
#          4     227.0
#          5     116.0
#          6     154.0
#          7     101.0
#          8     177.0
#          9     192.0
#          10    141.0
#          11     21.0