reptile
The first step is to start from natural crawler analyzes the request and page source code. From the web page source code, we can not find jobs posted. But in the request we see such a POST request
The figure below we can see that
-
url：https://www.lagou.com/jobs/positionAjax.json?city=%E5%8C%97%E4%BA%AC&needAddtionalResult=false
-
Request method: post
-
result: for the release of jobs
-
totalCount: as the number of jobs
Thinking
By analyzing the requests we find that 15 per return data, totalCount told us that the total number of job information.
Rounding up to the total number of pages to be acquired. Then save the resulting data to a csv file. So we get the data analysis of the data source!
- post request Form Data passed three parameters
- first: Are Home (and no use)
- pn: Page
- kd: Keyword Search
no bb， show code
'''
更多Python学习资料以及源码教程资料，可以在群821460695 免费获取
'''
# 获取请求结果
# kind 搜索关键字
# page 页码 默认是1
def get_json(kind, page=1,):
# post请求参数
param = {
'first': 'true',
'pn': page,
'kd': kind
}
header = {
'Host': 'www.lagou.com',
'Referer': 'https://www.lagou.com/jobs/list_python?labelWords=&fromSearch=true&suginput=',
'User-Agent': 'Mozilla/5.0 (Macintosh; Intel Mac OS X 10_12_6) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/69.0.3497.100 Safari/537.36'
}
# 设置代理
proxies = [
{'http': '140.143.96.216:80', 'https': '140.143.96.216:80'},
{'http': '119.27.177.169:80', 'https': '119.27.177.169:80'},
{'http': '221.7.255.168:8080', 'https': '221.7.255.168:8080'}
]
# 请求的url
url = 'https://www.lagou.com/jobs/positionAjax.json?px=default&city=%E5%8C%97%E4%BA%AC&needAddtionalResult=false'
# 使用代理访问
# response = requests.post(url, headers=header, data=param, proxies=random.choices(proxies))
response = requests.post(url, headers=header, data=param, proxies=proxies)
response.encoding = 'utf-8'
if response.status_code == 200:
response = response.json()
# 请求响应中的positionResult 包括查询总数 以及该页的招聘信息(公司名、地址、薪资、福利待遇等...)
return response['content']['positionResult']
return None
Then we just need to call get_json after each page request to obtain the results can be removed and then traverse the jobs needed
if __name__ == '__main__':
# 默认先查询第一页的数据
kind = 'python'
# 请求一次 获取总条数
position_result = get_json(kind=kind)
# 总条数
total = position_result['totalCount']
print('{}开发职位，招聘信息总共{}条.....'.format(kind, total))
# 每页15条 向上取整 算出总页数
page_total = math.ceil(total/15)
# 所有查询结果
search_job_result = []
#for i in range(1, total + 1)
# 为了节约效率 只爬去前100页的数据
for i in range(1, 100):
position_result = get_json(kind=kind, page= i)
# 每次抓取完成后,暂停一会,防止被服务器拉黑
time.sleep(15)
# 当前页的招聘信息
page_python_job = []
for j in position_result['result']:
python_job = []
# 公司全名
python_job.append(j['companyFullName'])
# 公司简称
python_job.append(j['companyShortName'])
# 公司规模
python_job.append(j['companySize'])
# 融资
python_job.append(j['financeStage'])
# 所属区域
python_job.append(j['district'])
# 职称
python_job.append(j['positionName'])
# 要求工作年限
python_job.append(j['workYear'])
# 招聘学历
python_job.append(j['education'])
# 薪资范围
python_job.append(j['salary'])
# 福利待遇
python_job.append(j['positionAdvantage'])
page_python_job.append(python_job)
# 放入所有的列表中
search_job_result += page_python_job
print('第{}页数据爬取完毕, 目前职位总数:{}'.format(i, len(search_job_result)))
# 每次抓取完成后,暂停一会,防止被服务器拉黑
time.sleep(15)
ok! We have to get the data, the last step we need to save the data in
# 将总数据转化为data frame再输出
df = pd.DataFrame(data=search_job_result,
columns=['公司全名', '公司简称', '公司规模', '融资阶段', '区域', '职位名称', '工作经验', '学历要求', '工资', '职位福利'])
df.to_csv('lagou.csv', index=False, encoding='utf-8_sig')
Run the main method directly on the results:
data analysis
In order to facilitate our statistics, we need to be cleaned by analyzing data cvs file
Excluding such recruitment trainee posts, without requiring or working life of the processing as a fresh raw 0, a salary range need to calculate the approximate value of as college education No
# 读取数据
df = pd.read_csv('lagou.csv', encoding='utf-8')
# 数据清洗,剔除实习岗位
df.drop(df[df['职位名称'].str.contains('实习')].index, inplace=True)
# print(df.describe())
# 由于CSV文件内的数据是字符串形式,先用正则表达式将字符串转化为列表,再取区间的均值
pattern = '\d+'
df['work_year'] = df['工作经验'].str.findall(pattern)
# 数据处理后的工作年限
avg_work_year = []
# 工作年限
for i in df['work_year']:
# 如果工作经验为'不限'或'应届毕业生',那么匹配值为空,工作年限为0
if len(i) == 0:
avg_work_year.append(0)
# 如果匹配值为一个数值,那么返回该数值
elif len(i) == 1:
avg_work_year.append(int(''.join(i)))
# 如果匹配值为一个区间,那么取平均值
else:
num_list = [int(j) for j in i]
avg_year = sum(num_list)/2
avg_work_year.append(avg_year)
df['工作经验'] = avg_work_year
# 将字符串转化为列表,再取区间的前25%，比较贴近现实
df['salary'] = df['工资'].str.findall(pattern)
# 月薪
avg_salary = []
for k in df['salary']:
int_list = [int(n) for n in k]
avg_wage = int_list[0]+(int_list[1]-int_list[0])/4
avg_salary.append(avg_wage)
df['月工资'] = avg_salary
# 将学历不限的职位要求认定为最低学历:大专\
df['学历要求'] = df['学历要求'].replace('不限','大专')
After the data through a simple cleaning, let's start our statistics
Draw a histogram salary
# 绘制频率直方图并保存
plt.hist(df['月工资'])
plt.xlabel('工资 (千元)')
plt.ylabel('频数')
plt.title("工资直方图")
plt.savefig('薪资.jpg')
plt.show()
Conclusion: Most of the salary Beijing Python development is between 15 ~ 25k
Locations pie chart
# 绘制饼图并保存
count = df['区域'].value_counts()
plt.pie(count, labels = count.keys(),labeldistance=1.4,autopct='%2.1f%%')
plt.axis('equal') # 使饼图为正圆形
plt.legend(loc='upper left', bbox_to_anchor=(-0.1, 1))
plt.savefig('pie_chart.jpg')
plt.show()
Conclusion: The most Python development company is Haidian District, followed by the Chaoyang District. Beijing ready to go to work the junior partner probably know where to rent it
Academic requirements histogram
# {'本科': 1304, '大专': 94, '硕士': 57, '博士': 1}
dict = {}
for i in df['学历要求']:
if i not in dict.keys():
dict[i] = 0
else:
dict[i] += 1
index = list(dict.keys())
print(index)
num = []
for i in index:
num.append(dict[i])
print(num)
plt.bar(left=index, height=num, width=0.5)
plt.show()
Conclusion: Python jobs, most companies require a bachelor's degree or higher. But education is just a stepping stone, if efforts to upgrade their technology, these are not the thing
'''
更多Python学习资料以及源码教程资料，可以在群821460695 免费获取
'''
# 绘制词云,将职位福利中的字符串汇总
text = ''
for line in df['职位福利']:
text += line
# 使用jieba模块将字符串分割为单词列表
cut_text = ' '.join(jieba.cut(text))
#color_mask = imread('cloud.jpg') #设置背景图
cloud = WordCloud(
background_color = 'white',
# 对中文操作必须指明字体
font_path='yahei.ttf',
#mask = color_mask,
max_words = 1000,
max_font_size = 100
).generate(cut_text)
# 保存词云图片
cloud.to_file('word_cloud.jpg')
plt.imshow(cloud)
plt.axis('off')
plt.show()
Conclusion: Flexible working is the company's most well-being, followed by five insurance payments few companies also provide six-insurance payments. Team atmosphere, flat management is also very important aspect.