import 
sys
 import 
pandas as 
pd
 import 
matplotlib.pyplot as 
plt
 import 
datetime as 
dt

 # 分割・併合データの読み込み 

bunkatsu_data = 
pd.read_csv( 'bunkatsu.dat' 
, sep= ' ' 
, names=[ 'code' 
, 'ds' 
, 'ratio' 
])
bunkatsu_data[ 'ds' 
] = 
bunkatsu_data[ 'ds' 
].str.replace( '/' 
, '-' 
)

 # 株価の読み取り 

code = 
9437 # 株価コード 

str_code = 
str(code)
filename = 
str_code + 
 '_1' 

read_data = 
pd.read_csv(filename, sep= ' ' 
, \
names=[ 'ds' 
, 'hajime' 
, 'takane' 
, 'yasune' 
, 'y' 
, 'dekidaka' 
, 'NaN' 
])
analysis_data = 
read_data.loc[:, [ 'y' 
, 'ds' 
]]
analysis_data[ 'ds' 
] = 
analysis_data[ 'ds' 
].str.replace( '/' 
, '-' 
)

 # 株価コードの分割データのみ取り出す 

bunkatsu_data = 
bunkatsu_data[bunkatsu_data[ 'code' 
] == 
code].reset_index(drop=True)

 # 分割データの適用 

 for 
i in 
range(len(bunkatsu_data)):
date = 
[int(s) for 
s in 
bunkatsu_data.at[i, 'ds' 
].split( '-' 
)] # 分割日
analysis_data.loc[pd.to_datetime(analysis_data[ 'ds' 
]) <= 
dt.datetime(date[0], date[1], date[2]), 'y' 
] \
 = 
analysis_data[ 'y' 
] / bunkatsu_data.at[i, 'ratio' 
]

 # プロット 

x = 
pd.to_datetime(analysis_data[ 'ds' 
])
fig = 
plt.figure(figsize=(20,5))
plt.plot(x, analysis_data[ 'y' 
][:],label= 'with split (' 
+str_code+ ')' 
)
plt.legend()
plt.show()
fig.savefig(str_code + 
 '.png' 
)

科学技術計算のためのPython入門ーー開発基礎，必須ライブラリ，高速化【電子書籍】[ 中久喜健司 ]

価格: 3,520円
(2020/8/19 09:34時点)
感想(0件)

#!/usr/bin/python3

 import 
sys
 from 
selenium import 
webdriver
 from 
bs4 import 
BeautifulSoup
 import 
requests

options = 
webdriver.ChromeOptions()
options.add_argument( '--headless' 
)
browser = 
webdriver.Chrome(options = 
options)

browser.get( '上記、カブコムのURL' 
)
http_src = 
browser.page_source
browser.quit()

rows = 
soup.find( 'tbody' 
).find_all( 'tr' 
)

f = 
open( 'bunkatsu.dat' 
, mode = 
 'w' 
)
 for 
row in 
rows:
cols = 
row.find_all( 'td' 
)
f.write(cols[1].text + 
 ' ' 
) # コード 

f.write(cols[4].text + 
 ' ' 
) # 分割適用前の最終日 

bunkatsu_data = 
cols[3].text.split( '：' 
)
f.write(str( float(bunkatsu_data[1])/float(bunkatsu_data[0]) ) + 
 '\n' 
) # 分割比 

f.close()

入門Python3 [ ビル・ルバノビック ]

価格: 4,070円
(2020/8/15 11:10時点)
感想(2件)

2006/1/4 1810 1822 1806 1815 74000
2006/1/5 1830 1834 1810 1815 314000
2006/1/6 1845 1859 1835 1847 317900
:
2020/8/5 5010 5040 4970 5020 171600
2020/8/6 4970 5010 4940 4960 160600
2020/8/7 4975 5050 4975 5010 119500

 import 
os
 import 
sys
 import 
pandas as 
pd
from fbprophet import 
Prophet
 import 
matplotlib.pyplot as 
plt

code = 
str(2331)
read_data = 
pd.read_csv( 'data/' 
+code, sep= ' ' 
, \
names=[ 'ds' 
, 'hajime' 
, 'takane' 
, 'yasune' 
, 'y' 
, 'dekidaka' 
, 'NaN' 
])

 # 直近500日ぐらいのデータを使用する 

analysis_data = 
read_data.loc[:, [ 'y' 
, 'ds' 
]].iloc[-500:, :].reset_index()
analysis_data[ 'ds' 
= analysis_data[ 'ds' 
].str.replace( '/' 
, '-' 
)

Maximum_value = 
analysis_data[ 'y' 
].max()
analysis_data[ 'cap' 
] = 
Maximum_value

model = 
Prophet(growth= 'logistic' 
)
model.fit(analysis_data)
future = 
model.make_future_dataframe(periods=60, freq = 
 'd' 
)
future[ 'cap' 
] = 
Maximum_value
future = 
future[ future[ 'ds' 
].dt.weekday < 5 ]
forecast = 
model.predict(future)

 # 図を作成 

fig = 
plt.figure(figsize=(20,5))
plt.plot(forecast[ 'yhat' 
][:],label= 'Prediction (' 
+code+ ')' 
)
plt.plot(analysis_data[ 'y' 
][:],label= 'True (' 
+code+ ')' 
)
plt.legend()
fig.savefig(code + 
 '.png' 
)

 # 直近30日の平均株価に対し、予想値の上昇率を計算 

Mean = 
analysis_data[ 'y' 
].iloc[-30:].mean()
Fct_value = 
forecast[ 'yhat' 
].iat[-1]
Rate = 
(Fct_value - Mean)/Mean*100
output_string = 
code+ ': ' 
\
+str(int(Mean))+ ' -> ' + 
str(int(Fct_value)) + 
' (' + 
str(int(Rate)) + 
'%)' 

if Rate > 10.0:
print( '!!! ' 
, output_string)
elif Rate > 5.0:
print( '!! ' 
, output_string)
else:
print(output_string)