python医疗大数据清洗_使用Python爬取、清洗并分析前程无忧的大数据职位

1 #-*- coding: utf-8 -*-

2 """

3 Created on Wed Nov 1 20:15:56 20194

5 @author: loo6 """

7

8 importmatplotlib.pyplot as plt9 importcsv10 importnumpy as np11 importre12 from wordcloud importWordCloud,STOPWORDS13

14

15 defreadFile():16 """

17 读取清洗后的文件18 """

19 data =[]20 with open("cleaned_51jobs.csv",encoding='gbk') as f:21 csv_reader = csv.reader(f) #使用csv.reader读取f中的文件

22 data_header = next(csv_reader) #读取第一行每一列的标题

23 for row in csv_reader: #将csv文件中的数据保存到data中

24 data.append(row)25

26 nd_data = np.array(data) #将list数组转化成array数组便于查看数据结构

27 jobName =nd_data[:, 0]28 locality = nd_data[:, 1]29 minSalary = nd_data[:, 2]30 maxSalary = nd_data[:, 3]31 returndata, jobName, locality, minSalary, maxSalary32

33

34