Python爬虫爬取电影票房数据及图表展示操作示例

import requests
import re
from matplotlib import pyplot as plt
from matplotlib import font_manager
import json

class DYOrder(object):
 #初始化
  def __init__(self,page=1):
    self.url = \'http://www.cbooo.cn/BoxOffice/getInland?pIndex={}&t=0\'.format(page)
    self.headers = {\'User-Agent\':\'Mozilla/5.0 (Macintosh; Intel Mac OS X 10_14_0) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/76.0.3809.100 Safari/537.36\'}
  #请求
  def __to_request(self):
    response = requests.get(url=self.url,headers=self.headers)
    return self.__to_parse(response.content.decode(\'utf-8\'))
  #解析
  def __to_parse(self,html):
    #返回为JSON字符串
    #首先将字符串反序列化为JSON对象
    my_json = json.loads(html)
    return my_json
  #图表展示
  def __to_show(self,data,show_type):
    x = []
    y = []
    for value in data:
      x.append(value[\'MovieName\'])
      y.append(int(value[\'BoxOffice\']))
    
    my_font = font_manager.FontProperties(fname=\'/System/Library/Fonts/PingFang.ttc\',size=18)
    
    if show_type == 1:
      plt.figure(figsize=(20,8),dpi=80)
      rects = plt.bar(range(len(x)),[float(i) for i in y],width=0.5,color=\'red\')
      plt.xticks(range(len(x)),x,fontproperties=my_font,rotation=60)
      plt.xlabel(\'名称\',rotation=60,color=\'blue\',fontproperties=my_font)
      plt.ylabel(\'票房/万\',rotation=60,color=\'blue\',fontproperties=my_font)
      for rect in rects:
        height = rect.get_height()
        plt.text(rect.get_x() + rect.get_width()/2,height+0.4,str(height),ha=\'center\',rotation=30)
    else:
      # 横向 plt.barh(y,x)
      plt.figure(figsize=(15,13),dpi=80)
      rects = plt.barh(range(len(x)),y,height=0.8,color=\'orange\')
      plt.yticks(range(len(x)),x,fontproperties=my_font,rotation=30)
      plt.ylabel(\'名称\',rotation=0,color=\'blue\',fontproperties=my_font)
      plt.xlabel(\'票房/万\',rotation=60,color=\'blue\',fontproperties=my_font)
      for rect in rects:
        width = rect.get_width()
        plt.text(width, rect.get_y()+0.3/2,str(width),va=\'center\',rotation=30)
  
    plt.grid(alpha=0.4)  
    plt.title(\'中国电影历史票房排行榜\',color=\'red\',size=18,fontproperties=my_font)
    plt.show()
  #所有操作
  def to_run(self,show_type=1):
    result = self.__to_request()
    self.__to_show(result,show_type)

if __name__ == \'__main__\':
  dy_order = DYOrder(1)
  # type 1 竖向条形图 2 横向
  dy_order.to_run(2)