数据分析实例（股票分析实例）

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

使用pandas的内建函数DataReader从雅虎财经网站读取股价数据

import tushare as ts

pingan = ts.get_k_data('601318',start='2011-01-01')

pingan.head()

	date	open	close	high	low	volume	code
0	2011-01-04	24.984	25.011	25.314	24.830	245626.0	601318
1	2011-01-05	24.870	24.110	24.962	24.083	427554.0	601318
2	2011-01-06	24.145	23.112	24.145	22.611	947078.0	601318
3	2011-01-07	23.279	23.297	23.842	22.765	659013.0	601318
4	2011-01-10	23.248	22.866	23.688	22.809	282919.0	601318

pingan.tail()

	date	open	close	high	low	volume	code
1999	2019-04-03	78.48	79.86	80.15	78.37	760219.0	601318
2000	2019-04-04	80.08	80.20	80.96	79.25	795626.0	601318
2001	2019-04-08	80.70	80.59	82.25	80.02	916122.0	601318
2002	2019-04-09	80.40	81.10	82.00	80.37	601737.0	601318
2003	2019-04-10	80.80	82.03	82.11	79.66	783620.0	601318

pingan['date'] = pd.to_datetime(pingan.date)

pingan.set_index('date',inplace = True)

pingan.index

DatetimeIndex(['2011-01-04', '2011-01-05', '2011-01-06', '2011-01-07',
               '2011-01-10', '2011-01-11', '2011-01-12', '2011-01-13',
               '2011-01-14', '2011-01-17',
               ...
               '2019-03-27', '2019-03-28', '2019-03-29', '2019-04-01',
               '2019-04-02', '2019-04-03', '2019-04-04', '2019-04-08',
               '2019-04-09', '2019-04-10'],
              dtype='datetime64[ns]', name='date', length=2004, freq=None)

pingan['close'].plot(figsize=(12,8))

g_yue = pingan['close'].resample('M').ohlc()
g_yue

	open	high	low	close
date
2011-01-31	25.011	25.011	21.249	21.842
2011-02-28	21.952	22.980	21.477	21.868
2011-03-31	22.084	23.512	21.433	21.736
2011-04-30	22.514	24.123	22.514	22.954
2011-05-31	23.103	23.103	20.972	21.407
2011-06-30	21.310	21.310	19.495	21.213
2011-07-31	21.156	21.609	19.534	19.724
2011-08-31	19.906	19.906	17.867	18.053
2011-09-30	18.231	18.294	14.935	14.935
2011-10-31	15.082	17.226	15.082	17.208
2011-11-30	17.449	18.036	15.687	15.687
2011-12-31	16.541	16.901	14.935	15.322
2012-01-31	15.082	17.395	15.060	17.057
2012-02-29	16.643	18.556	16.643	18.160
2012-03-31	18.040	18.352	16.123	16.274
2012-04-30	16.799	18.405	16.750	18.111
2012-05-31	18.761	19.166	18.085	18.636
2012-06-30	18.761	20.518	18.094	20.349
2012-07-31	20.509	20.572	19.493	19.793
2012-08-31	20.227	20.227	17.193	17.381
2012-09-30	17.829	18.833	17.560	18.833
2012-10-31	18.613	19.018	17.091	17.423
2012-11-30	17.661	17.661	15.865	16.382
2012-12-31	16.449	20.338	16.449	20.338
2013-01-31	21.155	21.653	19.952	21.653
2013-02-28	22.799	23.701	20.194	21.137
2013-03-31	20.634	20.634	18.245	18.757
2013-04-30	18.407	19.265	17.877	17.877
2013-05-31	17.976	18.717	17.374	17.920
2013-06-30	17.929	17.929	15.558	15.726
...	...	...	...	...
2016-11-30	33.393	35.031	32.964	34.602
2016-12-31	34.688	35.269	33.012	33.745
2017-01-31	33.983	34.774	33.536	34.774
2017-02-28	33.926	35.069	33.926	34.507
2017-03-31	34.440	35.250	33.907	35.250
2017-04-30	35.336	36.155	33.821	36.155
2017-05-31	36.107	43.327	35.488	42.927
2017-06-30	43.784	47.584	42.546	47.251
2017-07-31	46.613	52.003	45.641	50.087
2017-08-31	52.812	54.940	47.738	53.948
2017-09-30	54.083	54.685	51.567	52.616
2017-10-31	53.199	62.525	53.199	62.525
2017-11-30	62.292	76.203	62.185	67.926
2017-12-31	66.158	72.774	66.158	67.985
2018-01-31	70.520	76.514	68.014	72.939
2018-02-28	73.328	73.843	62.593	65.828
2018-03-31	66.965	71.977	63.448	63.448
2018-04-30	62.816	66.304	59.154	59.154
2018-05-31	59.591	62.272	58.921	60.358
2018-06-30	59.882	63.990	56.267	58.000
2018-07-31	54.673	61.545	54.673	60.990
2018-08-31	59.109	62.842	56.020	62.357
2018-09-30	62.109	68.500	60.600	68.500
2018-10-31	64.780	67.930	61.850	63.630
2018-11-30	64.300	67.350	62.640	63.490
2018-12-31	64.990	64.990	56.100	56.100
2019-01-31	55.180	62.980	55.180	62.980
2019-02-28	63.210	73.000	63.210	70.010
2019-03-31	72.390	77.100	68.700	77.100
2019-04-30	78.600	82.030	78.600	82.030

100 rows × 4 columns

pingan['close'].loc['2017'].plot(figsize=(12,8))

shouyi = (pingan['close'][-1] - pingan['close'][0])/pingan['close'][0]
shouyi

2.2797569069609374

y_num = pingan.index[-1].year - pingan.index[0].year
y_num

8

shouyi**(1/y_num)

1.1085009445718548

pingan['close'].to_period('A')

date
2011    25.011
2011    24.110
2011    23.112
2011    23.297
2011    22.866
2011    22.905
2011    23.288
2011    23.283
2011    22.844
2011    21.908
2011    21.974
2011    22.273
2011    21.438
2011    21.714
2011    21.249
2011    21.315
2011    21.504
2011    21.943
2011    21.710
2011    21.842
2011    21.952
2011    21.930
2011    22.224
2011    21.961
2011    22.980
2011    22.695
2011    22.629
2011    22.528
2011    22.492
2011    22.659
         ...  
2019    69.980
2019    70.010
2019    72.390
2019    72.580
2019    72.070
2019    72.160
2019    70.290
2019    68.700
2019    68.790
2019    69.250
2019    71.540
2019    72.800
2019    73.850
2019    75.000
2019    75.350
2019    76.550
2019    75.800
2019    74.720
2019    72.290
2019    72.690
2019    74.220
2019    73.400
2019    77.100
2019    78.600
2019    78.960
2019    79.860
2019    80.200
2019    80.590
2019    81.100
2019    82.030
Freq: A-DEC, Name: close, Length: 2004, dtype: float64

pingan['close'].to_period('A').groupby(level = 0).first().plot()

pingan['42d'] = pingan['close'].rolling(window=42).mean()
pingan['250d'] = pingan['close'].rolling(window=250).mean()

pingan[['close','42d','250d']].plot(figsize=(12,8),color = ['y','b','r'])

pingan['duishu'] = np.log(pingan['close']/pingan['close'].shift(1))   #shifting 指的是沿着时间轴将数据前移或后移。

pingan[['close','duishu']].plot(subplots = True,figsize=(12,12))

array([,
       ],
      dtype=object)