爱站程序员基地本期是对热播网剧《雪中悍刀行》的一次爬虫与数据分析，总爬取条数1W条评论，很适合新人练手。案例中还特别加入了对评论的情绪文本分析处理，供大家参考。
爬虫方面：由于腾讯的评论数据是封装在json里面，所以只需要找到json文件，对需要的数据进行提取保存即可。
视频网址：
https://v.qq.com/x/cover/mzc0020020cyvqh.html
评论json数据网址：
https://video.coral.qq.com/varticle/7579013546/comment/v2
注：只要替换视频数字id的值，即可爬取其他视频的评论
本文内容仅作为编程技术学习讨论，相关代码和数据不可用于商业用途，否则后果自负。
如何查找视频id？
通过Chrome开发者工具的Network查找请求。具体使用可参考旧文：爬虫必备工具，掌握它就解决了一半的问题
项目结构：
一. 爬虫部分：
1. 爬取评论内容代码：
spiders.py
import requestsimport reimport randomdef get_html(url, params):uapools = ['Mozilla/5.0 (Macintosh; Intel Mac OS X 10_9_2) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/35.0.1916.153 Safari/537.36','Mozilla/5.0 (Windows NT 6.1; WOW64; rv:30.0) Gecko/20100101 Firefox/30.0','Mozilla/5.0 (Macintosh; Intel Mac OS X 10_9_2) AppleWebKit/537.75.14 (19dffKHTML, like Gecko) Version/7.0.3 Safari/537.75.14']thisua = random.choice(uapools)headers = {"User-Agent": thisua}r = requests.get(url, headers=headers, params=params)r.raise_for_status()r.encoding = r.apparent_encodingr.encoding = 'utf-8'# 不加此句出现乱码return r.textdef parse_page(infolist, data):commentpat = '"content":"(.*?)"'lastpat = '"last":"(.*?)"'commentall = re.compile(commentpat, re.S).findall(data)next_cid = re.compile(lastpat).findall(data)[0]infolist.append(commentall)return next_ciddef print_comment_list(infolist):j = 0for page in infolist:print('第' + str(j + 1) + '页\\n')commentall = pagefor i in range(0, len(commentall)):print(commentall[i] + '\\n')j += 1def save_to_txt(infolist, path):fw = open(path, 'w+', encoding='utf-8')j = 0for page in infolist:#fw.write('第' + str(j + 1) + '页\\n')commentall = pagefor i in range(0, len(commentall)):fw.write(commentall[i] + '\\n')j += 1fw.close()def main():infolist = []vid = '7579013546';cid = "0";page_num = 3000url = 'https://video.coral.qq.com/varticle/' + vid + '/comment/v2'#print(url)for i in range(page_num):params = {'orinum': '10', 'cursor': cid}html = get_html(url, params)cid = parse_page(infolist, html)print_comment_list(infolist)save_to_txt(infolist, 'content.txt')main()
2. 爬取评论时间代码
sp.py
import requestsimport reimport randomdef get_html(url, params):uapools = ['Mozilla/5.0 (Macintosh; Intel Mac OS X 10_9_2) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/35.0.1916.153 Safari/537.36','Mozilla/5.0 (Windows NT 6.1; WOW64; rv:30.0) Gecko/20100101 Firefox/30.0','Mozilla/5.0 (Macintosh; Intel Mac OS X 10_9_2) AppleWebKit/537.75.14 (KHTML, like Gecko) Version/7.0.3 Safari/537.75.14']thisua = random.choice(uapools)headers = {"User-Agent": thisua}r = requests.get(url, headers=headers, params=params)r.raise_for_status()r.encoding = r.apparent_encodingr.encoding = 'utf-8'# 不加此句出现乱码return r.textdef parse_page(infolist, data):commentpat = '"time":"(.*?)"'lastpat = '"last":"(.*?)"'commentall = re.compile(commentpat, re.S).findall(data)next_cid = re.compile(lastpat).findall(data)[0]infolist.append(commentall)return next_ciddef print_comment_list(infolist):j = 0for page in infolist:print('第' + str(j + 1) + '页\\n')commentall = pagefor i in range(0, len(commentall)):print(commentall[i] + '\\n')j += 1def save_to_txt(infolist, path):fw = open(path, 'w+', encoding='utf-8')j = 0for page in infolist:#fw.write('第' + str(j + 1) + '页\\n')commentall = pagefor i in range(0, len(commentall)):fw.write(commentall[i] + '\\n')j += 1fw.close()def main():infolist = []vid = '7579013546';cid = "0";page_num =3000url = 'https://video.coral.qq.com/varticle/' + vid + '/comment/v2'#print(url)for i in range(page_num):params = {'orinum': '10', 'cursor': cid}html = get_html(url, params)cid = parse_page(infolist, html)print_comment_list(infolist)save_to_txt(infolist, 'time.txt')main()
二. 数据处理部分
1. 评论的时间戳转换为正常时间
time.py
# coding=gbkimport csvimport timecsvFile = open("data.csv",'w',newline='',encoding='utf-8')writer = csv.writer(csvFile)csvRow = []#print(csvRow)f = open("time.txt",'r',encoding='utf-8')for line in f:csvRow = int(line)#print(csvRow)timeArray = time.localtime(csvRow)csvRow = time.strftime("%Y-%m-%d %H:%M:%S", timeArray)print(csvRow)csvRow = csvRow.split()writer.writerow(csvRow)f.close()csvFile.close()
2. 评论内容读入csv
CD.py
# coding=gbkimport csvcsvFile = open("content.csv",'w',newline='',encoding='utf-8')writer = csv.writer(csvFile)csvRow = []f = open("content.txt",'r',encoding='utf-8')for line in f:csvRow = line.split()writer.writerow(csvRow)f.close()csvFile.close()
3. 统计一天各个时间段内的评论数
py.py
# coding=gbkimport csvfrom pyecharts import options as optsfrom sympy.combinatorics import Subsetfrom wordcloud import WordCloudwith open('../Spiders/data.csv') as csvfile:reader = csv.reader(csvfile)data1 = [str(row[1])[0:2] for row in reader]print(data1)print(type(data1))#先变成集合得到seq中的所有元素，避免重复遍历set_seq = set(data1)rst = []for item in set_seq:rst.append((item,data1.count(item))) #添加元素及出现个数rst.sort()print(type(rst))print(rst)with open("time2.csv", "w+", newline='', encoding='utf-8') as f:writer = csv.writer(f, delimiter=',')for i in rst: # 对于每一行的，将这一行的每个元素分别写在对应的列中writer.writerow(i)with open('time2.csv') as csvfile:reader = csv.reader(csvfile)x = [str(row[0]) for row in reader]print(x)with open('time2.csv') as csvfile:reader = csv.reader(csvfile)y1 = [float(row[1]) for row in reader]print(y1)
4. 统计最近评论数
py1.py
# coding=gbkimport csvfrom pyecharts import options as optsfrom sympy.combinatorics import Subsetfrom wordcloud import WordCloudwith open('../Spiders/data.csv') as csvfile:reader = csv.reader(csvfile)data1 = [str(row[0]) for row in reader]#print(data1)print(type(data1))#先变成集合得到seq中的所有元素，避免重复遍历set_seq = set(data1)rst = []for item in set_seq:rst.append((item,data1.count(item))) #添加元素及出现个数rst.sort()print(type(rst))print(rst)with open("time1.csv", "w+", newline='', encoding='utf-8') as f:writer = csv.writer(f, delimiter=',')for i in rst: # 对于每一行的，将这一行的每个元素分别写在对应的列中writer.writerow(i)with open('time1.csv') as csvfile:reader = csv.reader(csvfile)x = [str(row[0]) for row in reader]print(x)with open('time1.csv') as csvfile:reader = csv.reader(csvfile)y1 = [float(row[1]) for row in reader]print(y1)
三. 数据分析
数据分析方面：涉及到了词云图，条形，折线，饼图，后三者是对评论时间与主演占比的分析，然而腾讯的评论时间是以时间戳的形式显示，所以要进行转换，再去统计出现次数，最后，新加了对评论内容的情感分析。
1. 制作词云图
wc.py
import numpy as npimport reimport jiebafrom wordcloud import WordCloudfrom matplotlib import pyplot as pltfrom PIL import Image# 上面的包自己安装，不会的就百度f = open('content.txt', 'r', encoding='utf-8') # 这是数据源，也就是想生成词云的数据txt = f.read() # 读取文件f.close() # 关闭文件，其实用with就好，但是懒得改了# 如果是文章的话，需要用到jieba分词，分完之后也可以自己处理下再生成词云newtxt = re.sub("[A-Za-z0-9\\!\\%\\[\\]\\,\\。]", "", txt)print(newtxt)words = jieba.lcut(newtxt)img = Image.open(r'wc.jpg') # 想要搞得形状img_array = np.array(img)# 相关配置，里面这个collocations配置可以避免重复wordcloud = WordCloud(background_color="white",width=1080,height=960,font_path="../文悦新青年.otf",max_words=150,scale=10,#清晰度max_font_size=100,mask=img_array,collocations=False).generate(newtxt)plt.imshow(wordcloud)plt.axis('off')plt.show()wordcloud.to_file('wc.png')
轮廓图：wc.jpg
词云图：result.png （注：这里把英文字母过滤掉）
2. 制作最近评论数条形图
DrawBar.py
# encoding: utf-8import csvimport pyecharts.options as optsfrom pyecharts.charts import Barfrom pyecharts.globals import ThemeTypeclass DrawBar(object):"""绘制柱形图类"""def __init__(self):"""创建柱状图实例，并设置宽高和风格"""self.bar = Bar(init_opts=opts.InitOpts(width='1500px', height='700px', theme=ThemeType.LIGHT))def add_x(self):"""为图形添加X轴数据"""with open('time1.csv') as csvfile:reader = csv.reader(csvfile)x = [str(row[0]) for row in reader]print(x)self.bar.add_xaxis(xaxis_data=x,)def add_y(self):with open('time1.csv') as csvfile:reader = csv.reader(csvfile)y1 = [float(row[1]) for row in reader]print(y1)"""为图形添加Y轴数据，可添加多条"""self.bar.add_yaxis( # 第一个Y轴数据series_name="评论数", # Y轴数据名称y_axis=y1, # Y轴数据label_opts=opts.LabelOpts(is_show=True,color="black"), # 设置标签bar_max_width='100px', # 设置柱子最大宽度)def set_global(self):"""设置图形的全局属性"""#self.bar(width=2000,height=1000)self.bar.set_global_opts(title_opts=opts.TitleOpts( # 设置标题title='雪中悍刀行近日评论统计',title_textstyle_opts=opts.TextStyleOpts(font_size=35)),tooltip_opts=opts.TooltipOpts( # 提示框配置项（鼠标移到图形上时显示的东西）is_show=True, # 是否显示提示框trigger="axis", # 触发类型（axis坐标轴触发，鼠标移到时会有一条垂直于X轴的实线跟随鼠标移动，并显示提示信息）axis_pointer_type="cross"# 指示器类型（cross将会生成两条分别垂直于X轴和Y轴的虚线，不启用trigger才会显示完全）),toolbox_opts=opts.ToolboxOpts(), # 工具箱配置项(什么都不填默认开启所有工具))def draw(self):"""绘制图形"""self.add_x()self.add_y()self.set_global()self.bar.render('../Html/DrawBar.html') # 将图绘制到 test.html 文件内，可在浏览器打开def run(self):"""执行函数"""self.draw()if __name__ == '__main__':app = DrawBar()app.run()
效果图：DrawBar.html
3. 制作每小时评论条形图
DrawBar2.py
# encoding: utf-8# encoding: utf-8import csvimport pyecharts.options as optsfrom pyecharts.charts import Barfrom pyecharts.globals import ThemeTypeclass DrawBar(object):"""绘制柱形图类"""def __init__(self):"""创建柱状图实例，并设置宽高和风格"""self.bar = Bar(init_opts=opts.InitOpts(width='1500px', height='700px', theme=ThemeType.MACARONS))def add_x(self):"""为图形添加X轴数据"""str_name1 = '点'with open('time2.csv') as csvfile:reader = csv.reader(csvfile)x = [str(row[0] + str_name1) for row in reader]print(x)self.bar.add_xaxis(xaxis_data=x)def add_y(self):with open('time2.csv') as csvfile:reader = csv.reader(csvfile)y1 = [int(row[1]) for row in reader]print(y1)"""为图形添加Y轴数据，可添加多条"""self.bar.add_yaxis( # 第一个Y轴数据series_name="评论数", # Y轴数据名称y_axis=y1, # Y轴数据label_opts=opts.LabelOpts(is_show=False), # 设置标签bar_max_width='50px', # 设置柱子最大宽度)def set_global(self):"""设置图形的全局属性"""#self.bar(width=2000,height=1000)self.bar.set_global_opts(title_opts=opts.TitleOpts( # 设置标题title='雪中悍刀行各时间段评论统计',title_textstyle_opts=opts.TextStyleOpts(font_size=35)),tooltip_opts=opts.TooltipOpts( # 提示框配置项（鼠标移到图形上时显示的东西）is_show=True, # 是否显示提示框trigger="axis", # 触发类型（axis坐标轴触发，鼠标移到时会有一条垂直于X轴的实线跟随鼠标移动，并显示提示信息）axis_pointer_type="cross"# 指示器类型（cross将会生成两条分别垂直于X轴和Y轴的虚线，不启用trigger才会显示完全）),toolbox_opts=opts.ToolboxOpts(), # 工具箱配置项(什么都不填默认开启所有工具))def draw(self):"""绘制图形"""self.add_x()self.add_y()self.set_global()self.bar.render('../Html/DrawBar2.html') # 将图绘制到 test.html 文件内，可在浏览器打开def run(self):"""执行函数"""self.draw()if __name__ == '__main__':app = DrawBar()app.run()
效果图：DrawBar2.html
4. 制作近日评论数饼图
pie_pyecharts.py
import csvfrom pyecharts import options as optsfrom pyecharts.charts import Piefrom random import randintfrom pyecharts.globals import ThemeTypewith open('time1.csv') as csvfile:reader = csv.reader(csvfile)x = [str(row[0]) for row in reader]print(x)with open('time1.csv') as csvfile:reader = csv.reader(csvfile)y1 = [float(row[1]) for row in reader]print(y1)num = y1lab = x(Pie(init_opts=opts.InitOpts(width='1700px',height='450px',theme=ThemeType.LIGHT))#默认900，600.set_global_opts(title_opts=opts.TitleOpts(title="雪中悍刀行近日评论统计",title_textstyle_opts=opts.TextStyleOpts(font_size=27)),legend_opts=opts.LegendOpts(pos_top="10%", pos_left="1%",# 图例位置调整),).add(series_name='',center=[280, 270], data_pair=[(j, i) for i, j in zip(num, lab)])#饼图.add(series_name='',center=[845, 270],data_pair=[(j,i) for i,j in zip(num,lab)],radius=['40%','75%'])#环图.add(series_name='', center=[1380, 270],data_pair=[(j, i) for i, j in zip(num, lab)], rosetype='radius')#南丁格尔图).render('pie_pyecharts.html')
效果图
5. 制作每小时评论饼图
pie_pyecharts2.py
import csvfrom pyecharts import options as optsfrom pyecharts.charts import Piefrom random import randintfrom pyecharts.globals import ThemeType
str_name1 = '点'with open('time2.csv') as csvfile:reader = csv.reader(csvfile)x = [str(row[0]+str_name1) for row in reader]print(x)with open('time2.csv') as csvfile:reader = csv.reader(csvfile)y1 = [int(row[1]) for row in reader]print(y1)num = y1lab = x(Pie(init_opts=opts.InitOpts(width='1650px',height='500px',theme=ThemeType.LIGHT,))#默认900，600.set_global_opts(title_opts=opts.TitleOpts(title="雪中悍刀行每小时评论统计",title_textstyle_opts=opts.TextStyleOpts(font_size=27)),legend_opts=opts.LegendOpts(pos_top="8%", pos_left="4%",# 图例位置调整),).add(series_name='',center=[250, 300], data_pair=[(j, i) for i, j in zip(num, lab)])#饼图.add(series_name='',center=[810, 300],data_pair=[(j,i) for i,j in zip(num,lab)],radius=['40%','75%'])#环图.add(series_name='', center=[1350, 300],data_pair=[(j, i) for i, j in zip(num, lab)], rosetype='radius')#南丁格尔图).render('pie_pyecharts2.html')
效果图
6. 制作观看时间区间评论统计饼图
pie_pyecharts3.py
# coding=gbkimport csvfrom pyecharts import options as optsfrom pyecharts.globals import ThemeTypefrom sympy.combinatorics import Subsetfrom wordcloud import WordCloudfrom pyecharts.charts import Piefrom random import randint
with open(/data.csv') as csvfile:reader = csv.reader(csvfile)data2 = [int(row[1].strip('')[0:2]) for row in reader]#print(data2)print(type(data2))#先变成集合得到seq中的所有元素，避免重复遍历set_seq = set(data2)list = []for item in set_seq:list.append((item,data2.count(item))) #添加元素及出现个数list.sort()print(type(list))#print(list)with open("time2.csv", "w+", newline='', encoding='utf-8') as f:writer = csv.writer(f, delimiter=',')for i in list: # 对于每一行的，将这一行的每个元素分别写在对应的列中writer.writerow(i)n = 4#分成n组m = int(len(list)/n)list2 = []for i in range(0, len(list), m):list2.append(list[i:i+m])print("凌晨 : ",list2[0])print("上午 : ",list2[1])print("下午 : ",list2[2])print("晚上 : ",list2[3])with open('time2.csv') as csvfile:reader = csv.reader(csvfile)y1 = [int(row[1]) for row in reader]print(y1)n =6groups = [y1[i:i + n] for i in range(0, len(y1), n)]print(groups)x=['凌晨','上午','下午','晚上']y1=[]for y1 in groups:num_sum = 0for groups in y1:num_sum += groupsstr_name1 = '点'num = y1lab = x(Pie(init_opts=opts.InitOpts(width='1500px',height='450px',theme=ThemeType.LIGHT))#默认900，600.set_global_opts(title_opts=opts.TitleOpts(title="雪中悍刀行观看时间区间评论统计", title_textstyle_opts=opts.TextStyleOpts(font_size=30)),legend_opts=opts.LegendOpts(pos_top="8%", # 图例位置调整),).add(series_name='',center=[260, 270], data_pair=[(j, i) for i, j in zip(num, lab)])#饼图.add(series_name='',center=[1230, 270],data_pair=[(j,i) for i,j in zip(num,lab)],radius=['40%','75%'])#环图.add(series_name='', center=[750, 270],data_pair=[(j, i) for i, j in zip(num, lab)], rosetype='radius')#南丁格尔图).render('pie_pyecharts3.html')
效果图
7. 制作雪中悍刀行主演提及占比饼图
pie_pyecharts4.py
import csvfrom pyecharts import options as optsfrom pyecharts.charts import Piefrom random import randintfrom pyecharts.globals import ThemeTypef = open('content.txt', 'r', encoding='utf-8') # 这是数据源，也就是想生成词云的数据words = f.read() # 读取文件f.close() # 关闭文件，其实用with就好，但是懒得改了name=["张若昀","李庚希","胡军"]print(name)count=[float(words.count("张若昀")),float(words.count("李庚希")),float(words.count("胡军"))]print(count)num = countlab = name(Pie(init_opts=opts.InitOpts(width='1650px',height='450px',theme=ThemeType.LIGHT))#默认900，600.set_global_opts(title_opts=opts.TitleOpts(title="雪中悍刀行主演提及占比",title_textstyle_opts=opts.TextStyleOpts(font_size=27)),legend_opts=opts.LegendOpts(pos_top="3%", pos_left="33%",# 图例位置调整),).add(series_name='',center=[280, 270], data_pair=[(j, i) for i, j in zip(num, lab)])#饼图.add(series_name='',center=[800, 270],data_pair=[(j,i) for i,j in zip(num,lab)],radius=['40%','75%'])#环图.add(series_name='', center=[1300, 270],data_pair=[(j, i) for i, j in zip(num, lab)], rosetype='radius')#南丁格尔图).render('pie_pyecharts4.html')
效果图
8. 评论内容情感分析
SnowNLP.py
import numpy as npfrom snownlp import SnowNLPimport matplotlib.pyplot as pltf = open('content.txt', 'r', encoding='UTF-8')list = f.readlines()sentimentslist = []for i in list:s = SnowNLP(i)print(s.sentiments)sentimentslist.append(s.sentiments)plt.hist(sentimentslist, bins=np.arange(0, 1, 0.01), facecolor='g')plt.xlabel('Sentiments Probability')plt.ylabel('Quantity')plt.title('Analysis of Sentiments')plt.show()
效果图（情感各分数段出现频率）
SnowNLP情感分析是基于情感词典实现的，其简单的将文本分为两类，积极和消极，返回值为情绪的概率，也就是情感评分在[0,1]之间，越接近1，情感表现越积极，越接近0，情感表现越消极。
总结
以上就是如何获取腾讯视频评论并整理后进行可视化分析，属于数据分析及可视化的常规操作，没有特别高深的技术，还比较适合数据分析初学者练习，读者可以自行动手尝试。全部代码已在文中给出。
