数据分析基础

import numpy as np
import pandas as pd
import scipy
import matplotlib.pyplot as plt
'''数据导入'''
df = pd.read_csv(r'C:UsersAdministratorDesktopapp01_bgm.csv',encoding='gbk')
print(df.head())#前五行
print(df.tail())#最后五行
print(df.columns[[1,2]])#头信息
print(df.index)
print(df.T)#数据转置使用T方法，
print(df.ix[:,0].head())
print(df.ix[10:20,1])#行索引，列索引
print(df.drop(df.columns[[3]],axis=1))#axis 参数告诉函数到底舍弃列还是行。如果axis等于0，那么就舍弃行。
print(df.describe())#通过describe属性，对数据的统计特性进行描述
'''
               id          fen           num
count  2216.00000  2216.000000   2216.000000
mean   2519.50000     6.755415    618.274819
std     639.84842     0.842402    959.487784
min    1412.00000     1.200000     51.000000
25%    1965.75000     6.200000    112.000000
50%    2519.50000     6.800000    253.000000
75%    3073.25000     7.300000    693.250000
max    3627.00000     9.100000  10112.000000
'''