資料視覺化

抓資料

import requests

from io import StringIO

import pandas as pd

import numpy as np

import datetime

import time

def crawl_price(date):

r = requests.post('http://www.twse.com.tw/exchangeReport/MI_INDEX?response=csv&date=' + str(date).split(' ')[0].replace('-','') + '&type=ALL')

ret = pd.read_csv(StringIO(r.text.replace("=", "")), header=["證券代號" in l for l in r.text.split("\n")].index(True)-1)

ret = ret.set_index('證券代號')

ret['成交金額'] = ret['成交金額'].str.replace(',','')

ret['成交股數'] = ret['成交股數'].str.replace(',','')

return ret

data = {}

n_days = 60

date = datetime.datetime.now()

fail_count = 0

allow_continuous_fail_count = 5

while len(data) < n_days:

print('parsing', date)

# 使用 crawl_price 抓資料

try:

# 抓資料

data[date.date()] = crawl_price(date)

print('success!')

fail_count = 0

except:

# 假日爬不到

print('fail! check the date is holiday')

fail_count += 1

if fail_count == allow_continuous_fail_count:

raise

break

# 減一天

date -= datetime.timedelta(days=1)

time.sleep(10)

close = pd.DataFrame({k:d['收盤價'] for k,d in data.items()}).transpose()

close.index = pd.to_datetime(close.index)

open = pd.DataFrame({k:d['開盤價'] for k,d in data.items()}).transpose()

open.index = pd.to_datetime(open.index)

high = pd.DataFrame({k:d['最高價'] for k,d in data.items()}).transpose()

high.index = pd.to_datetime(high.index)

low = pd.DataFrame({k:d['最低價'] for k,d in data.items()}).transpose()

low.index = pd.to_datetime(low.index)

volume = pd.DataFrame({k:d['成交股數'] for k,d in data.items()}).transpose()

volume.index = pd.to_datetime(volume.index)

tsmc = {

'close':close['2330']['2020'].dropna().astype(float),

'open':open['2330']['2020'].dropna().astype(float),

'high':high['2330']['2020'].dropna().astype(float),

'low':low['2330']['2020'].dropna().astype(float),

'volume': volume['2330']['2020'].dropna().astype(float),

}

tsmc['close'].plot()