[英](Python, DataFrame): Record the average of all numbers in a column that are smaller than the n'th percentile
I have a DataFrame similar to the below and would like to create a DataFrame or series that looks more like the second table. 我有一个类似于下面的DataFrame,并想创建一个看起来更像第二张表的DataFrame或系列。
For example: I would find the nth percentile of column A, then take the average of all numbers in A that are less than the nth percentile. 例如:我将找到列A的第n个百分位数,然后取A中小于第n个百分位数的所有数字的平均值。
I've used the code below to get the average and range of each column but seem to be missing something to get the conditional average. 我使用下面的代码来获取每列的平均值和范围,但似乎缺少一些东西来获取条件平均值。
min = df.min(axis='index')
max = df.max(axis='index')
mean = df.mean(axis = 'index')
df[df < np.percentile(df, 0.4)].mean()
this doesnt seem to work and I believe gives the average of every row 这似乎不起作用,我相信可以给出每一行的平均值
Table 1 表格1
Date A B C D E F
02/10/2017 10 5 1 2 1 1
01/10/2017 10 4 9 4 3 5
30/09/2017 4 8 5 6 2 4
29/09/2017 8 2 7 9 10 5
28/09/2017 3 8 2 7 10 8
27/09/2017 7 3 8 9 9 7
26/09/2017 4 1 2 9 3 4
25/09/2017 10 1 6 6 3 5
24/09/2017 8 3 5 5 6 7
23/09/2017 7 9 5 7 1 3
22/09/2017 2 9 10 5 8 1
Table 2 表2
Index Avg<40th Percentile
A 3.25
B 1.333333333
C 1.666666667
D 4
E 1.333333333
F 1.666666667
Use 采用
df.where(df < df.quantile(0.4)).mean()
Date NaN
A 3.250000
B 1.333333
C 1.666667
D 4.000000
E 1.333333
F 1.666667
声明:本站的技术帖子网页,遵循CC BY-SA 4.0协议,如果您需要转载,请注明本站网址或者原文地址。任何问题请咨询:yoyou2525@163.com.