哪吒數據提取、數據分析
- 2019 年 10 月 5 日
- 筆記
版權聲明：本文為部落客原創文章，遵循 CC 4.0 BY-SA 版權協議，轉載請附上原文出處鏈接和本聲明。
本文鏈接：https://blog.csdn.net/weixin_43908900/article/details/100882598
最近哪吒大火，所以我們分析一波哪吒的影評資訊，分析之前我們需要數據呀，所以開篇我們先講一下爬蟲的數據提取；話不多說，走著。
首先我們找到網站的url = "https://maoyan.com/films/1211270",找到評論區看看網友的吐槽，如下
F12打開看看有沒有評論資訊，我們發現還是有資訊的。
但是現在的問題時，我們好像只有這幾條評論資訊，完全不支援我們的分析呀，我們只能另謀出路了；
f12中由手機測試功能，打開刷新頁面，向下滾動看見查看好幾十萬的評論數據，點擊進入後，在network中會看見url = "http://m.maoyan.com/review/v2/comments.json?movieId=1211270&userId=-1&offset=15&limit=15&ts=1568600356382&type=3"api,有這個的時候我們就可以搞事情了。
但是隨著爬取，還是不能獲取完整的資訊，百度、Google、必應一下，我們通過時間段獲取資訊，這樣我們不會被貓眼給牆掉，所以我們使用該 url="http://m.maoyan.com/mmdb/comments/movie/1211270.json?_v_=yes&offset=0&startTime="
效果如下：
開始構造爬蟲程式碼：
#!/usr/bin/env python # -*- coding: utf-8 -*- # author：albert time:2019/9/3 import requests,json,time,csv from fake_useragent import UserAgent #獲取userAgent from datetime import datetime,timedelta def get_content(url): '''獲取api資訊的網頁源程式碼''' ua = UserAgent().random try: data = requests.get(url,headers={'User-Agent':ua},timeout=3 ).text return data except: pass def Process_data(html): '''對數據內容的獲取''' data_set_list = [] #json格式化 data_list = json.loads(html)['cmts'] for data in data_list: data_set = [data['id'],data['nickName'],data['userLevel'],data['cityName'],data['content'],data['score'],data['startTime']] data_set_list.append(data_set) return data_set_list if __name__ == '__main__': start_time = start_time = datetime.now().strftime('%Y-%m-%d %H:%M:%S') # 獲取當前時間，從當前時間向前獲取 # print(start_time) end_time = '2019-07-26 08:00:00' # print(end_time) while start_time > str(end_time): #構造url url = 'http://m.maoyan.com/mmdb/comments/movie/1211270.json?_v_=yes&offset=0&startTime=' + start_time.replace( ' ', '%20') print('........') try: html = get_content(url) except Exception as e: time.sleep(0.5) html = get_content(url) else: time.sleep(1) comments = Process_data(html) # print(comments[14][-1]) if comments: start_time = comments[14][-1] start_time = datetime.strptime(start_time, '%Y-%m-%d %H:%M:%S') + timedelta(seconds=-1) # print(start_time) start_time = datetime.strftime(start_time,'%Y-%m-%d %H:%M:%S') print(comments) #保存數據為csv with open("comments_1.csv", "a", encoding='utf-8',newline='') as csvfile: writer = csv.writer(csvfile) writer.writerows(comments)
———————————–數據分析部分———————————–
我們手裡有接近兩萬的數據後開始進行數據分析階段：
工具：jupyter、庫方法：pyecharts v1.0===> pyecharts 庫向下不兼容，所以我們需要使用新的方式（鏈式結構）實現：
我們先來分析一下哪吒的等級星圖，使用pandas 實現分組求和，正對1-5星的數據：
from pyecharts import options as opts from pyecharts.globals import SymbolType from pyecharts.charts import Bar,Pie,Page,WordCloud from pyecharts.globals import ThemeType,SymbolType import numpy import pandas as pd df = pd.read_csv('comments_1.csv',names=["id","nickName","userLevel","cityName","score","startTime"]) attr = ["一星", "二星", "三星", "四星", "五星"] score = df.groupby("score").size() # 分組求和 value = [ score.iloc[0] + score.iloc[1]+score.iloc[1], score.iloc[3] + score.iloc[4], score.iloc[5] + score.iloc[6], score.iloc[7] + score.iloc[8], score.iloc[9] + score.iloc[10], ]
# 餅圖分析 # 暫時處理，不能直接調用value中的數據 attr = ["一星", "二星", "三星", "四星", "五星"] value = [286, 43, 175, 764, 10101] pie = ( Pie(init_opts=opts.InitOpts(theme=ThemeType.LIGHT)) .add('',[list(z) for z in zip(attr, value)]) .set_global_opts(title_opts=opts.TitleOpts(title='哪吒等級分析')) .set_series_opts(label_opts=opts.LabelOpts(formatter="{b}:{c}")) ) pie.render_notebook()
實現效果：
然後進行詞雲分析：
import jieba import matplotlib.pyplot as plt #生成圖形 from wordcloud import WordCloud,STOPWORDS,ImageColorGenerator df = pd.read_csv("comments_1.csv",names =["id","nickName","userLevel","cityName","content","score","startTime"]) comments = df["content"].tolist() # comments df # 設置分詞 comment_after_split = jieba.cut(str(comments), cut_all=False) # 非全模式分詞，cut_all=false words = " ".join(comment_after_split) # 以空格進行拼接 stopwords = STOPWORDS.copy() stopwords.update({"電影","最後","就是","不過","這個","一個","感覺","這部","雖然","不是","真的","覺得","還是","但是"}) bg_image = plt.imread('bg.jpg') #生成 wc=WordCloud( width=1024, height=768, background_color="white", max_words=200, mask=bg_image, #設置圖片的背景 stopwords=stopwords, max_font_size=200, random_state=50, font_path='C:/Windows/Fonts/simkai.ttf' #中文處理，用系統自帶的字體 ).generate(words) #產生背景圖片，基於彩色影像的顏色生成器 image_colors=ImageColorGenerator(bg_image) #開始畫圖 plt.imshow(wc.recolor(color_func=image_colors)) #為背景圖去掉坐標軸 plt.axis("off") #保存雲圖 plt.show() wc.to_file("哪吒.png")
效果如下：
