时间序列数据的采样和画图

时间序列数据的采样和画图

引入相关库

import numpy as np
import pandas as pd
from pandas import Series,DataFrame

时间序列数据的采样

使用date_range创建一个datetime对象，从2016-01-01开始，以天为间隔，一共366天

t_range=pd.date_range('2016-01-01','2016-12-31')
t_range

DatetimeIndex(['2016-01-01', '2016-01-02', '2016-01-03', '2016-01-04',
               '2016-01-05', '2016-01-06', '2016-01-07', '2016-01-08',
               '2016-01-09', '2016-01-10',
               ...
               '2016-12-22', '2016-12-23', '2016-12-24', '2016-12-25',
               '2016-12-26', '2016-12-27', '2016-12-28', '2016-12-29',
               '2016-12-30', '2016-12-31'],
              dtype='datetime64[ns]', length=366, freq='D')

以该datetime对象作为index，创建一个Series，长度也为366

s1=Series(np.random.randn(len(t_range)),index=t_range)
s1

2016-01-01    1.406461
2016-01-02    0.447931
2016-01-03    1.699394
2016-01-04    0.788757
2016-01-05   -0.232023
                ...   
2016-12-27   -0.247121
2016-12-28    0.673201
2016-12-29    0.091558
2016-12-30   -1.163763
2016-12-31    0.482879
Freq: D, Length: 366, dtype: float64

s1

2016-01-01    1.406461
2016-01-02    0.447931
2016-01-03    1.699394
2016-01-04    0.788757
2016-01-05   -0.232023
                ...   
2016-12-27   -0.247121
2016-12-28    0.673201
2016-12-29    0.091558
2016-12-30   -1.163763
2016-12-31    0.482879
Freq: D, Length: 366, dtype: float64

s1['2016-01']

2016-01-01    1.406461
2016-01-02    0.447931
2016-01-03    1.699394
2016-01-04    0.788757
2016-01-05   -0.232023
2016-01-06    0.348905
2016-01-07   -0.236980
2016-01-08    0.161807
2016-01-09    0.218637
2016-01-10    1.672460
2016-01-11    0.983214
2016-01-12    0.044190
2016-01-13   -0.620090
2016-01-14    0.101304
2016-01-15    1.801919
2016-01-16    1.624366
2016-01-17    0.672093
2016-01-18    1.475754
2016-01-19    0.582916
2016-01-20   -0.284506
2016-01-21   -1.634910
2016-01-22    0.000536
2016-01-23   -2.254585
2016-01-24    0.261707
2016-01-25   -0.320744
2016-01-26   -0.173244
2016-01-27    0.544120
2016-01-28    0.059852
2016-01-29    0.479251
2016-01-30    0.857616
2016-01-31   -0.815376
Freq: D, dtype: float64

s1['2016-01'].mean()

0.3116365177410832

把该Series采样，变成十二个月的数据，可以通过mean来算取每个月的评估值，一共十二个月，得到采样Series的大小为12，采样方法可以使用resample方法为按月进行采样。

s1_month=s1.resample('M').mean()
s1_month

2016-01-31    0.311637
2016-02-29   -0.055687
2016-03-31    0.152305
2016-04-30    0.058371
2016-05-31   -0.102482
2016-06-30   -0.164970
2016-07-31   -0.298555
2016-08-31    0.041087
2016-09-30    0.216605
2016-10-31    0.137500
2016-11-30   -0.269811
2016-12-31   -0.038291
Freq: M, dtype: float64

s1_month.index

DatetimeIndex(['2016-01-31', '2016-02-29', '2016-03-31', '2016-04-30',
               '2016-05-31', '2016-06-30', '2016-07-31', '2016-08-31',
               '2016-09-30', '2016-10-31', '2016-11-30', '2016-12-31'],
              dtype='datetime64[ns]', freq='M')

把resample方法改成按小时采样，由于原始数据一天只有一个数据点，所以变成按小时采样需要填充数据填满24个小时，ffill填充方法为使用一天的这个数据点往前填充，填充24小时

s1.resample('H').ffill()

2016-01-01 00:00:00    1.406461
2016-01-01 01:00:00    1.406461
2016-01-01 02:00:00    1.406461
2016-01-01 03:00:00    1.406461
2016-01-01 04:00:00    1.406461
                         ...   
2016-12-30 20:00:00   -1.163763
2016-12-30 21:00:00   -1.163763
2016-12-30 22:00:00   -1.163763
2016-12-30 23:00:00   -1.163763
2016-12-31 00:00:00    0.482879
Freq: H, Length: 8761, dtype: float64

bfill方法的填充数据来自下一天即2016-01-02的数据

s1.resample('H').bfill()

2016-01-01 00:00:00    1.406461
2016-01-01 01:00:00    0.447931
2016-01-01 02:00:00    0.447931
2016-01-01 03:00:00    0.447931
2016-01-01 04:00:00    0.447931
                         ...   
2016-12-30 20:00:00    0.482879
2016-12-30 21:00:00    0.482879
2016-12-30 22:00:00    0.482879
2016-12-30 23:00:00    0.482879
2016-12-31 00:00:00    0.482879
Freq: H, Length: 8761, dtype: float64

时间序列数据的画图

使用date_range再创建一个datetime，从2016-01-01到2016-12-31以每小时为单位，长度为8761

t_range=pd.date_range('2016-01-01','2016-12-31',freq='H')
t_range

DatetimeIndex(['2016-01-01 00:00:00', '2016-01-01 01:00:00',
               '2016-01-01 02:00:00', '2016-01-01 03:00:00',
               '2016-01-01 04:00:00', '2016-01-01 05:00:00',
               '2016-01-01 06:00:00', '2016-01-01 07:00:00',
               '2016-01-01 08:00:00', '2016-01-01 09:00:00',
               ...
               '2016-12-30 15:00:00', '2016-12-30 16:00:00',
               '2016-12-30 17:00:00', '2016-12-30 18:00:00',
               '2016-12-30 19:00:00', '2016-12-30 20:00:00',
               '2016-12-30 21:00:00', '2016-12-30 22:00:00',
               '2016-12-30 23:00:00', '2016-12-31 00:00:00'],
              dtype='datetime64[ns]', length=8761, freq='H')

创建一个空的DataFrame，以datetime作为index

stock_df=DataFrame(index=t_range)
stock_df.head()

2016-01-01 00:00:00
2016-01-01 01:00:00
2016-01-01 02:00:00
2016-01-01 03:00:00
2016-01-01 04:00:00

创建一个columns，作为阿里巴巴的股票，从数值范围为80-160，长度为8761

stock_df['BABA']=np.random.randint(80,160,size=len(t_range))
stock_df.head()

	BABA
2016-01-01 00:00:00	84
2016-01-01 01:00:00	80
2016-01-01 02:00:00	111
2016-01-01 03:00:00	159
2016-01-01 04:00:00	94

创建一个腾讯的股票，数值范围为30-50

stock_df['TENCENT']=np.random.randint(30,50,size=len(t_range))
stock_df.head()

	BABA	TENCENT
2016-01-01 00:00:00	84	33
2016-01-01 01:00:00	80	42
2016-01-01 02:00:00	111	48
2016-01-01 03:00:00	159	39
2016-01-01 04:00:00	94	49

对这个两列的DataFrame画图，创建一个matplotlib里面的图，这个图蓝线表示阿里，黄线表示腾讯，但是由于点太多比较拥挤

stock_df.plot()

<matplotlib.axes._subplots.AxesSubplot at 0x19e61fe4dc8>

如果上述方法显示不出图片的解决方法

import matplotlib.pyplot as plt
plt.show()

对DataFrame重新采样，变成每周一个点

weekly_df=DataFrame()

对阿里巴巴的值按每周重新采样

weekly_df['BABA']=stock_df['BABA'].resample('W').mean()

对腾讯的值按每周重新采样

weekly_df['TENCENT']=stock_df['TENCENT'].resample('W').mean()

weekly_df.head()

	BABA	TENCENT
2016-01-03	117.708333	39.916667
2016-01-10	120.476190	39.767857
2016-01-17	122.428571	39.636905
2016-01-24	116.702381	39.297619
2016-01-31	120.708333	39.839286

展示曲线，可以清楚的看到两条线

weekly_df.plot()

<matplotlib.axes._subplots.AxesSubplot at 0x19e629e5bc8>

上面方法不能显示的解决办法

plt.show()