Question

我要按数据框在“ zone_id和eventName”两列上进行分组。我需要计算按zone_id分组的eventName的百分比。

换句话说，我需要通过zone_id计算（点击/打印）* 100。

import pandas as pd

#read the csv file
df = pd.read_csv('data.csv', sep=';')

result=df.groupby(['zone_id','eventName']).event.count()

print(result)

#I use count() method to extract the number of clicked and printed by zone_id. Then on this basis I think to be able to find a way to compute a     percentage by zone_id.

output : 
zone_id  eventName
28       printed         88
9283     clicked         197
         printed         7732
9284     clicked         2
         printed         452
9287     clicked         129
         printed         3802
9614     clicked         4
         printed         342
17437    clicked         55
         printed         4026

#By using mean() function, the mean calculation is well done grouped by zone_id
result=df.groupby(['zone_id','eventName']).event.count().groupby('zone_id').mean()

print(result)

output :
zone_id
28         88.0
9283     3964.5
9284      227.0
9287     1965.5
9614      173.0
17437    2040.5

#Expected result : I need to compute the percentage of eventName (clicked/printed)*100 by zone_id
 Expected output:
zone_id
28        0%    -> (0/88)*100
9283      2.54% -> (197/7732)*100
9284      0.44% -> (2/452)*100
9287      3.39% -> (129/3802)*100
9614      1.16% -> (4/342)*100
17437     1.36% -> (55/4026)*100

Answer 1

没有示例数据很难看到，但是尝试这样的事情吗？

events = df.groupby(['zone_id','eventName']).size()
events.loc[pd.IndexSlice[:, 'printed']] / events.loc[pd.IndexSlice[:, 'clicked']]

或者使用unstack获取点击并打印为列：

events = df.groupby(['zone_id','eventName']).size().unstack(level=1)
events['printed'] / events['clicked']

如何在按两列分组的数据框中计算百分比

1 个答案: