最近想学做数据分析,可是干啃书实在太无聊,看着看着就走神了。所以干脆找点有意思的事情做。碰巧看到知乎上有人教怎么做词频统计还有图片https://www.zhihu.com/question/28975391/answer/100796070
又突然想起之前看到有人爬咪蒙的文章做统计,我想了想干脆我也做咪蒙的词频统计吧。
首先,我得先找到咪蒙的所有文章。由于周末在家,我用的是六年前的老笔记本,实在太卡。我也不打算爬她的公众号了。
这个我百度一下去找现成的她的文章。
http://www.360doc.com/content/16/1026/12/34624569_601474728.shtml
这次我用scrapy框架来爬取,上代码(不懂scrapy的童鞋请自行看官方文档学习基础知识)
spider
import scrapy
from bs4 import BeautifulSoup
from ali.items import AliItem
class ali(scrapy.Spider):
name = 'aliyun'
n=0
def start_requests(self):
start_url=['http://www.360doc.com/content/16/1026/12/34624569_601474728.shtml']
for url in start_url:
yield scrapy.Request(url=url,callback=self.parse_getlink)
def parse_getlink(self,response):
n=0
#print response.url
soup = BeautifulSoup(response.body,'lxml')
for i in soup.find_all('td',id='artContent'):#定位元素,拿到每篇文章的链接
for j in i.find_all('a'):
url= j.get('href')
yield scrapy.Request(url=url, callback=self.parse)
n=n+1
print n
def parse(self,response):
item = AliItem()
soup = BeautifulSoup(response.body,'lxml')
for i in soup.find_all('div',id='js_content'):#定位元素,拿到每篇文章的内容
item['content'] = i.get_text()
yield item#把内容传到pipeline上去处理
爬虫的逻辑写完了,然后我们到pipeline
去处理爬回来的内容
# -*- coding: utf-8 -*-
# Define your item pipelines here
#
# Don't forget to add your pipeline to the ITEM_PIPELINES setting
# See: http://doc.scrapy.org/en/latest/topics/item-pipeline.html
class AliPipeline(object):
def process_item(self, item, spider):#把爬回来的文章内容写到mimeng.txt上去
f=open('mimeng.txt','a')
content = item['content'].encode('utf-8')
f.writelines(content)
f.close()
return item
QQ截图20161224211635.png
然后我们就要对文章内容进行分词了!这里用到jieba这个库,没错就是结巴。
写个新的py
import jieba
f = open('mimeng.txt', 'r',encoding='utf-8').read()
w = open('mimengcount.txt','w')
words = list(jieba.cut(f))#用jieba分词
for word in words:
if len(word) > 1:
word = word + '\n'
#a.append(word)
w.writelines(word)
w.close()
QQ截图20161224211842.png
接下来就要用四个库来进行词频统计和作画
numpy,wordcloud,PLS以及matplotlib
下面这段代码我是复制别人的,其实我也不懂那几个库哈哈
import jieba
from wordcloud import WordCloud
import PIL
import matplotlib.pyplot as plt
import numpy as np
def wordcloudplot():
text = open('mimengcount.txt').read()
path='ziti.ttf'#字体文件,没有这个中文会乱码,自己去百度下载
#path=unicode(path,'utf8').encode('gb18030')
alice_mask = np.array(PIL.Image.open('3e2f.jpg'))
wordcloud = WordCloud(font_path=path,background_color="white", margin=5, width=1800, height=800, mask=alice_mask, max_words=2000,
max_font_size=60, random_state=42)
worcloud = wordcloud.generate(text)
wordcloud.to_file('pic3.jpg')
plt.imshow(wordcloud)
plt.axis('off')
plt.show()
wordcloudplot()
下面,我们来见证奇迹
pic3.jpg看到这结果。。。。我突然有种浪费了几小时什么的感觉。。。本来今天下了王者荣耀想要去玩的。。。我特么的犯贱跑去做这个。。。真无聊
网友评论