In [83]:

    

# -*- coding: utf-8 -*-
import matplotlib.pyplot as plt
import pandas as pd
plt.rcParams['font.sans-serif']=['SimHei'] #用来正常显示中文标签
plt.rcParams['axes.unicode_minus']=False #用来正常显示负号
#有中文出现的情况，需要u'内容'

df = pd.read_excel(u'用户标签.xlsx',sheetname=u'标签', encoding = "gbk",index_col=1,header=1)
df.shape


    

    
Out[83]:





(25555, 58)

In [105]:

    

df.dtypes
# X[2:,2]
df.head
# df[df[u"Unnamed: 1"]<>Nil].head
# df = df.loc[:,['Unnamed: 1',u'娱乐偏好']]
df.columns.size
df.iloc[:,1]
fig = plt.figure()
ax = fig.add_subplot(331)
# ax.hist(df[u'消费水平'])
plt.title(u'消费水平分布')
plt.xlabel(u'消费水平')
plt.ylabel(u'人数')
plt.show()

# df = df.iloc[:,1]
var = df.fillna('nil').groupby(u'消费水平')[u'消费水平'].count()
ax1 = fig.add_subplot(332)
ax1.set_xlabel(u'消费水平')
ax1.set_ylabel(u'人数')
ax1.set_title(u'消费水平分布2')
var.plot(kind='bar')
plt.show()
# df.describe
var2 = df.fillna('nil').groupby(u'购物')[u'购物'].count()
ax2 = fig.add_subplot(334)
ax2.set_xlabel(u'购物')
ax2.set_ylabel(u'人数')
ax2.set_title(u'购物')
var2.plot(kind='bar')
plt.show()
var3 = df.fillna('nil').groupby(u'新闻')[u'新闻'].count()
ax3 = fig.add_subplot(334)
ax3.set_xlabel(u'新闻')
ax3.set_ylabel(u'人数')
ax3.set_title(u'新闻')
var3.plot(kind='bar')
plt.show()


    

In [ ]: