基于OPGG的数据分析

 1 # _*_ coding:utf-8 _*_
 2 # from 坾尘
 3 # 2020/5/09
 4 from selenium import webdriver
 5 import time
 6 from bs4 import BeautifulSoup
 7 import matplotlib.pyplot as plt
 8 plt.rcParams[‘font.sans-serif‘] = [‘SimHei‘]
 9 driver =webdriver.PhantomJS(executable_path="phantomjs.exe")
10 driver.get("http://www.op.gg/champion/statistics")
11 time.sleep(5)
12 file=open(‘opgggg.csv‘,‘w‘,encoding=‘utf-8‘)
13 demo=driver.page_source
14 soup=BeautifulSoup(demo,‘lxml‘)
15 
16 b,f,c=[],[],[]
17 a=soup.find_all(name=[‘td‘],class_=[‘champion-index-table__cell champion-index-table__cell--champion‘])
18 for i in a :
19     b.append(i.a.div.text)
20 d=soup.find_all(name=[‘td‘],class_=[‘champion-index-table__cell champion-index-table__cell--value‘])
21 for i in d:
22     if i.text!=‘\n\n‘:
23         f.append(float(i.text.replace(‘%‘,‘‘)))
24 file.write(‘英雄名称‘+‘,‘+‘胜率‘+‘,‘+‘登场率‘+‘\n‘)
25 for i in range(len(b)):
26     file.write(str(b[i])+‘,‘+str(f[2*i-1])+‘,‘+str(f[2*i])+‘\n‘)
27 plt.ylim([45,55])
28 plt.title(u"op英雄")
29 plt.ylabel(u"胜率%")
30 plt.xlabel(u"英雄")
31 plt.bar(b[:10],f[:20:2],width = 0.35,align=‘center‘,color = ‘c‘,alpha=0.8)
32 plt.show()
33 print(b)
34 print(f)

driver.get("http://www.op.gg/champion/statistics")
time.sleep(5)

demo=driver.page_source