pandas知识点汇总

pandas基础知识汇总

---

1.时间序列

import pandas as pd
import numpy as np
import matplotlib.pyplot as plt

from datetime import datetime

now=datetime.now()
now

datetime.datetime(2018, 11, 18, 16, 44, 4, 405600)

print(now.strftime('%Y-%m-%d'))
print(datetime.strptime('7/6/2018','%m/%d/%Y'))
print(now.strftime('%X'))

2018-11-18
2018-07-06 00:00:00
16:44:04

dates=pd.date_range('11/1/2018',periods=50,freq='W-WED')
long_df=pd.DataFrame(np.random.randn(50,4),index=dates,columns=list('ABCD'))
long_df.head(10)

.dataframe tbody tr th:only-of-type {
vertical-align: middle;
}

.dataframe tbody tr th {
    vertical-align: top;
}

.dataframe thead th {
    text-align: right;
}

	A	B	C	D
2018-11-07	0.215536	0.855986	0.737170	-0.440150
2018-11-14	-0.477099	0.467430	-0.107105	0.941922
2018-11-21	0.052926	-0.671084	0.219058	-0.350776
2018-11-28	-1.449668	0.003958	1.065875	-0.277673
2018-12-05	1.371631	0.542839	0.071466	0.609508
2018-12-12	0.322176	1.335534	-0.423240	-0.111549
2018-12-19	-0.564089	0.262918	0.477552	0.018652
2018-12-26	-0.490212	0.382492	-0.858712	-0.920786
2019-01-02	1.630409	-0.740542	1.296362	0.376437
2019-01-09	1.460070	-0.449293	-0.783725	-1.098911

resample=long_df.resample('M').mean()
resample

.dataframe tbody tr th:only-of-type {
vertical-align: middle;
}

.dataframe tbody tr th {
    vertical-align: top;
}

.dataframe thead th {
    text-align: right;
}

	A	B	C	D
2018-11-30	-0.414576	0.164073	0.478750	-0.031669
2018-12-31	0.159876	0.630946	-0.183234	-0.101044
2019-01-31	0.092189	-0.225606	0.251072	-0.456075
2019-02-28	-0.124615	-0.467522	-0.142258	0.195602
2019-03-31	-0.294693	-0.014264	0.725285	1.291576
2019-04-30	0.182648	0.231022	-0.458572	0.294329
2019-05-31	0.317648	0.060677	0.297406	-0.035691
2019-06-30	0.407404	-0.198072	-0.461785	1.074969
2019-07-31	-0.245908	0.150161	0.526564	-0.082258
2019-08-31	0.046819	-0.227364	-0.684359	0.033979
2019-09-30	-0.834454	1.186670	0.653583	-0.306585
2019-10-31	-0.436990	-0.460347	0.040175	0.681903

pd.date_range('11/18/2018',periods=10,freq='2h30min')

DatetimeIndex(['2018-11-18 00:00:00', '2018-11-18 02:30:00',
               '2018-11-18 05:00:00', '2018-11-18 07:30:00',
               '2018-11-18 10:00:00', '2018-11-18 12:30:00',
               '2018-11-18 15:00:00', '2018-11-18 17:30:00',
               '2018-11-18 20:00:00', '2018-11-18 22:30:00'],
              dtype='datetime64[ns]', freq='150T')

type(resample)

pandas.core.resample.DatetimeIndexResampler

ts=pd.Series(np.arange(10),index=pd.date_range('11/18/2018',periods=10,freq='T'))
ts

2018-11-18 00:00:00    0
2018-11-18 00:01:00    1
2018-11-18 00:02:00    2
2018-11-18 00:03:00    3
2018-11-18 00:04:00    4
2018-11-18 00:05:00    5
2018-11-18 00:06:00    6
2018-11-18 00:07:00    7
2018-11-18 00:08:00    8
2018-11-18 00:09:00    9
Freq: T, dtype: int32

#pay attention to the parameter 'closed'
ts.resample('3min',closed='left',label='left').sum()

2018-11-18 00:00:00     3
2018-11-18 00:03:00    12
2018-11-18 00:06:00    21
2018-11-18 00:09:00     9
Freq: 3T, dtype: int32

ts.resample('3min').ohlc()

.dataframe tbody tr th:only-of-type {
vertical-align: middle;
}

.dataframe tbody tr th {
    vertical-align: top;
}

.dataframe thead th {
    text-align: right;
}

	open	high	low	close
2018-11-18 00:00:00	0	2	0	2
2018-11-18 00:03:00	3	5	3	5
2018-11-18 00:06:00	6	8	6	8
2018-11-18 00:09:00	9	9	9	9

long_df.plot()

## 滑窗函数
fig,axes=plt.subplots(1,3,figsize=(20,4))
long_df['A'].plot(ax=axes[0])
long_df['A'].rolling(window=10).mean().plot(ax=axes[0],title='A_10_mean')
long_df['B'].plot(ax=axes[1])
long_df['B'].rolling(window=10).sum().plot(ax=axes[1],title='B_10_sum')
long_df['C'].plot(ax=axes[2])
long_df['C'].rolling(window=10).quantile(quantile=0.8).plot(ax=axes[2],title='C_10_quantile')

#corr
from pylab import mpl
mpl.rcParams['font.sans-serif'] = ['SimHei'] # 指定默认字体
mpl.rcParams['axes.unicode_minus'] = False # 解决保存图像是负号'-'显示为方块的问题
long_df['B'].rolling(window=10).corr(long_df['A']).plot(style='ro--',grid=True,title='二元函数相关系数')

2.matplotlib绘图

long_df['A'].plot(kind='kde',style='g')

pd.plotting.scatter_matrix(long_df,diagonal='kde',color='r')

df=pd.DataFrame(np.random.randn(6,4),index='one two three four five six'.split(' '),columns=list('ABCD'))
df_normal=abs(df).div(abs(df).sum(1),axis=0)
df_normal.plot(kind='barh',stacked=True)
abs(df).sum(1)

one      3.989060
two      1.160160
three    2.087209
four     2.680116
five     4.452365
six      2.298789
dtype: float64