pandas-时序数据的位移与调整

最新推荐文章于 2023-05-06 01:25:53 发布

妮酱也爱敲代码

最新推荐文章于 2023-05-06 01:25:53 发布

阅读量286

点赞数 1

分类专栏：小象学院文章标签：数据分析

本文链接：https://blog.youkuaiyun.com/lildn/article/details/115032662

版权

小象学院专栏收录该内容

58 篇文章

订阅专栏

本文详细介绍了Pandas库中用于处理时序数据位移和调整的asfreq API，通过该方法可以将数据转换为特定频率的时间序列。链接：https://pandas.pydata.org/pandasdocs/stable/reference/api/pandas.DataFrame.asfreq.html。

摘要生成于 C知道，由 DeepSeek-R1 满血版支持，前往体验 >

asfreq的API（地址：https://pandas.pydata.org/pandasdocs/stable/reference/api/pandas.DataFrame.asfreq.html）
在这里插入图片描述

第四课 Pandas时序型数据分析
第四节 时序数据的位移与频率调整
import pandas as pd
import matplotlib.pyplot as plt

%matplotlib inline
#parse_dates=True将时间序列解析为datetime类型
data_df = pd.read_csv('./datasets/day_stats.csv', index_col='date', parse_dates=True, dayfirst=True)
data_df.head()
city	PM_China	PM_US Post	Polluted State CH	Polluted State US
date					
2013-10-01	beijing	67.416667	71.458333	light	light
2013-10-10	beijing	74.041667	81.583333	light	medium
2013-10-11	beijing	59.694444	59.291667	light	light
2013-10-12	beijing	69.236111	70.458333	light	light
2013-10-13	beijing	54.513889	57.333333	light	light
beijing_data = data_df[data_df['city'] == 'beijing'].copy()
beijing_data['PM_China'].plot(figsize=(10, 5))
<matplotlib.axes._subplots.AxesSubplot at 0x1de1cc83ba8>

位移 shift
beijing_data['+1'] = beijing_data['PM_China'].shift(1)
beijing_data['-1'] = beijing_data['PM_China'].shift(-1)
beijing_data.head()
city	PM_China	PM_US Post	Polluted State CH	Polluted State US	+1	-1
date							
2013-10-01	beijing	67.416667	71.458333	light	light	NaN	74.041667
2013-10-10	beijing	74.041667	81.583333	light	medium	67.416667	59.694444
2013-10-11	beijing	59.694444	59.291667	light	light	74.041667	69.236111
2013-10-12	beijing	69.236111	70.458333	light	light	59.694444	54.513889
2013-10-13	beijing	54.513889	57.333333	light	light	69.236111	37.739130
beijing_data.loc['2015-01'][['PM_China', '+1', '-1']].plot(figsize=(10, 5))
<matplotlib.axes._subplots.AxesSubplot at 0x1de1cdab8d0>

频率调整 asfreq
# 高频转低频
beijing_data.asfreq(freq='M').head()
city	PM_China	PM_US Post	Polluted State CH	Polluted State US	+1	-1
date							
2013-10-31	beijing	142.666667	170.750000	medium	heavy	92.583333	160.055556
2013-11-30	beijing	34.166667	34.708333	good	good	41.763889	44.565217
2013-12-31	beijing	49.722222	49.333333	light	light	44.571429	107.402778
2014-01-31	beijing	142.680556	166.208333	medium	heavy	65.930556	140.222222
2014-02-28	beijing	84.194444	102.666667	medium	medium	16.651515	6.304348
beijing_data['2013-12-31']
city	PM_China	PM_US Post	Polluted State CH	Polluted State US	+1	-1
date							
2013-12-31	beijing	49.722222	49.333333	light	light	44.571429	107.402778

# 低频转高频
beijing_data.asfreq(freq='12H').head(20)
city	PM_China	PM_US Post	Polluted State CH	Polluted State US	+1	-1
date							
2013-10-01 00:00:00	beijing	67.416667	71.458333	light	light	NaN	74.041667
2013-10-01 12:00:00	NaN	NaN	NaN	NaN	NaN	NaN	NaN
2013-10-02 00:00:00	beijing	16.680556	19.541667	good	good	60.944444	24.971014
2013-10-02 12:00:00	NaN	NaN	NaN	NaN	NaN	NaN	NaN
2013-10-03 00:00:00	beijing	62.041667	61.041667	light	light	31.500000	92.583333
2013-10-03 12:00:00	NaN	NaN	NaN	NaN	NaN	NaN	NaN
2013-10-04 00:00:00	beijing	160.055556	168.166667	heavy	heavy	142.666667	286.782609
2013-10-04 12:00:00	NaN	NaN	NaN	NaN	NaN	NaN	NaN
2013-10-05 00:00:00	beijing	286.782609	305.913044	heavy	heavy	160.055556	219.152778
2013-10-05 12:00:00	NaN	NaN	NaN	NaN	NaN	NaN	NaN
2013-10-06 00:00:00	beijing	219.152778	213.416667	heavy	heavy	286.782609	105.472222
2013-10-06 12:00:00	NaN	NaN	NaN	NaN	NaN	NaN	NaN
2013-10-07 00:00:00	beijing	105.472222	91.291667	medium	medium	219.152778	78.129630
2013-10-07 12:00:00	NaN	NaN	NaN	NaN	NaN	NaN	NaN
2013-10-08 00:00:00	beijing	78.129630	63.611111	medium	light	105.472222	145.541667
2013-10-08 12:00:00	NaN	NaN	NaN	NaN	NaN	NaN	NaN
2013-10-09 00:00:00	beijing	145.541667	146.500000	medium	medium	78.129630	204.305556
2013-10-09 12:00:00	NaN	NaN	NaN	NaN	NaN	NaN	NaN
2013-10-10 00:00:00	beijing	74.041667	81.583333	light	medium	67.416667	59.694444
2013-10-10 12:00:00	NaN	NaN	NaN	NaN	NaN	NaN	NaN

# 低频转高频，使用fill_value填充空值
beijing_data.asfreq(freq='12H', fill_value=0).head(20)
city	PM_China	PM_US Post	Polluted State CH	Polluted State US	+1	-1
date							
2013-10-01 00:00:00	beijing	67.416667	71.458333	light	light	NaN	74.041667
2013-10-01 12:00:00	0	0.000000	0.000000	0	0	0.000000	0.000000
2013-10-02 00:00:00	beijing	16.680556	19.541667	good	good	60.944444	24.971014
2013-10-02 12:00:00	0	0.000000	0.000000	0	0	0.000000	0.000000
2013-10-03 00:00:00	beijing	62.041667	61.041667	light	light	31.500000	92.583333
2013-10-03 12:00:00	0	0.000000	0.000000	0	0	0.000000	0.000000
2013-10-04 00:00:00	beijing	160.055556	168.166667	heavy	heavy	142.666667	286.782609
2013-10-04 12:00:00	0	0.000000	0.000000	0	0	0.000000	0.000000
2013-10-05 00:00:00	beijing	286.782609	305.913044	heavy	heavy	160.055556	219.152778
2013-10-05 12:00:00	0	0.000000	0.000000	0	0	0.000000	0.000000
2013-10-06 00:00:00	beijing	219.152778	213.416667	heavy	heavy	286.782609	105.472222
2013-10-06 12:00:00	0	0.000000	0.000000	0	0	0.000000	0.000000
2013-10-07 00:00:00	beijing	105.472222	91.291667	medium	medium	219.152778	78.129630
2013-10-07 12:00:00	0	0.000000	0.000000	0	0	0.000000	0.000000
2013-10-08 00:00:00	beijing	78.129630	63.611111	medium	light	105.472222	145.541667
2013-10-08 12:00:00	0	0.000000	0.000000	0	0	0.000000	0.000000
2013-10-09 00:00:00	beijing	145.541667	146.500000	medium	medium	78.129630	204.305556
2013-10-09 12:00:00	0	0.000000	0.000000	0	0	0.000000	0.000000
2013-10-10 00:00:00	beijing	74.041667	81.583333	light	medium	67.416667	59.694444
2013-10-10 12:00:00	0	0.000000	0.000000	0	0	0.000000	0.000000