pandas分区间,算频率的实例

import pandas as pd

path='F:/python/python数据分析与挖掘实战/图书配套数据、代码/chapter3/demo/data/catering_fish_congee.xls'

data=pd.read_excel(path,header=None,index_col=0)

data.index.name='日期'

data.columns=['销售额(元)']

  

xse=data['销售额(元)']

print(xse.max())

print(xse.min())

print(xse.max()-xse.min())

  

fanwei=list(range(0,4500,500))

fenzu=pd.cut(xse.values,fanwei,right=False)#分组区间,长度91

print(fenzu.codes)#标签

print(fenzu.categories)#分组区间，长度8

pinshu=fenzu.value_counts()#series,区间-个数

print(pinshu.index)

  

import matplotlib.pyplot as plt

pinshu.plot(kind='bar')

#plt.text(0,29,str(29))

  

  

qujian=pd.cut(xse,fanwei,right=False)

data['区间']=qujian.values

data.groupby('区间').median()

data.groupby('区间').mean()#每个区间平均数

  

pinshu_df=pd.DataFrame(pinshu,columns=['频数'])

pinshu_df['频率f']=pinshu_df / pinshu_df['频数'].sum()

pinshu_df['频率%']=pinshu_df['频率f'].map(lambda x:'%.2f%%'%(x*100))

  

pinshu_df['累计频率f']=pinshu_df['频率f'].cumsum()

pinshu_df['累计频率%']=pinshu_df['累计频率f'].map(lambda x:'%.4f%%'%(x*100))

  

In[158]: pinshu_df

Out[158]:

       频数    频率f   频率%   累计频率f   累计频率%

[0, 500)   29 0.318681 31.87% 0.318681  31.8681%

[500, 1000)  20 0.219780 21.98% 0.538462  53.8462%

[1000, 1500) 12 0.131868 13.19% 0.670330  67.0330%

[1500, 2000) 12 0.131868 13.19% 0.802198  80.2198%

[2000, 2500)  8 0.087912  8.79% 0.890110  89.0110%

[2500, 3000)  3 0.032967  3.30% 0.923077  92.3077%

[3000, 3500)  4 0.043956  4.40% 0.967033  96.7033%

[3500, 4000)  3 0.032967  3.30% 1.000000 100.0000%