数据缺失主要包括记录缺失和字段信息缺失等情况，其对数据分析会有较大影响，导致结果不确定性更加显著。

缺失值填充2种思路：

均值/中位数/众数填充、临近值填充

均值/中位数/众数填充

s = pd.Series([1,np.nan,np.nan,2,3,np.nan,4])#print(s)print('------')
# 分别求出均值/中位数/众数u = s.mean()     # 均值me = s.median()  # 中位数mod = s.mode()   # 众数print('均值为：%.2f, 中位数为：%.2f' % (u,me))print('众数为：', mod.tolist())print('------')
# 用均值填补s.fillna(u,inplace = True)print(s)

------
均值为：2.50, 中位数为：2.50
众数为：[1.0, 2.0, 3.0, 4.0]
------
0    1.0
1    2.5
2    2.5
3    2.0
4    3.0
5    2.5
6    4.0
dtype: float64

临近值填充：向下填充/向上填充

s = pd.Series([1,np.nan,np.nan,2,3,np.nan,4])#print(s)print('------')
# 用前值插补/向下填充s.fillna(method = 'ffill',inplace = True)print(s)

s = pd.Series([1,np.nan,np.nan,2,3,np.nan,4])#print(s)print('------')
# 用后值插补/向上填充s.fillna(method = 'bfill',inplace = True)print(s)

------
0    1.0
1    1.0
2    1.0
3    2.0
4    3.0
5    3.0
6    4.0
dtype: float64
------
0    1.0
1    2.0
2    2.0
3    2.0
4    3.0
5    4.0
6    4.0
dtype: float64

缺失值填充实例：密度图​​​​​​​

data = pd.Series(np.random.rand(100)*100)data[3,6,33,45,56,66,67,80,90] = np.nanprint(data.head())print('总数据量:%i' % len(data))print('------')
data_na = data[data.isnull()]print('缺失值数据量:%i' % len(data_na))print('缺失数据占比:%.2f%%' % (len(data_na) / len(data) * 100))

# 密度图查看缺失值情况data_c = data.fillna(data.median())  #  中位数填充缺失值data_f = data.fillna(method = 'ffill')  #  前值填充缺失值fig,axes = plt.subplots(1,4,figsize = (10,6))data.plot.box(ax = axes[0],grid = True,title = '数据分布')data.plot(kind = 'kde',style = '--r',ax = axes[1],grid = True,title = '删除缺失值',xlim = [-50,150])data_c.plot(kind = 'kde',style = '--b',ax = axes[2],grid = True,title = '中位数填充缺失值',xlim = [-50,150])data_f.plot(kind = 'kde',style = '--g',ax = axes[3],grid = True,title = '前值填充缺失值',xlim = [-50,150])

0    39.117918
1     1.340284
2    95.345331
3          NaN
4    13.084512
dtype: float64
总数据量:100
------
缺失值数据量:9
缺失数据占比:9.00%