python数据分析入门

秒速五厘米 2022-04-03 04:58 308阅读 0赞

python数据分析入门，作为入门文章系列主要包含以下几个内容：

1.数据的来源（本案例采用的数据来自于上一篇文章中爬取的智联招聘信息）：读取数据库数据、数据写入csv文件、读取csv文件等

2.数据的处理：对载入到内存的数据进行一系列的操作（处理总共包含清洗、过滤、分组、统计、排序、计算平均数等一系列操作，本文只简单涉及了其中某几个）

3.对处理后的数据进行可视化分析等

# ！/usr/bin/env python
    # -*-coding:utf-8-*-
    """
    @Author  : xiaofeng
    @Time    : 2018/12/19 15:23
    @Desc : Less interests,More interest.（数据分析入门）
    @Project : python_appliction
    @FileName: analysis1.py
    @Software: PyCharm
    @Blog    ：https://blog.csdn.net/zwx19921215
    """
    import pymysql as db
    import pandas as pd
    import numpy as np
    import seaborn as sns
    import matplotlib as mpl
    import matplotlib.pyplot as plt
    
    # mysql config
    mysql_config = {
        'host': '110.0.2.130',
        'user': 'test',
        'password': 'test',
        'database': 'xiaofeng',
        'charset': 'utf8'
    }
    
    """
    read data from mysql and write to local file
    （从数据库中读取数据写入本地文件）
    @:param page_no
    @:param page_size
    @:param path 文件写入路径
    """
    
    
    def read_to_csv(page_no, page_size, path):
        # read from database
        database = db.connect(**mysql_config)
        if page_no > 1:
            page_no = (page_no - 1) * page_size
        else:
            page_no = 0
        sql = 'select * from zhilian limit ' + str(page_no) + ',' + str(page_size) + ''
        df = pd.read_sql(sql, database)
        print(df)
        database.close()
        # write to csv
        local = path
        df.to_csv(local, encoding='gbk')
    
    
    """
    read data from remote address or local address
    （读取文件从远程地址或者本地）
    """
    
    
    def read_from_csv(path):
        # remote address
        # data_url = 'https://xxxx/zhilian.csv'
        # local address
        # data_url = 'G:/zhilian.csv'
        df = pd.read_csv(path, encoding='gbk')
        return df
    
    
    """
    数据集的简单（过滤、分组、排序）处理
    """
    
    
    def simple_op(path):
        df = read_from_csv(path)
        # top 10 ：获取前10行
        print('--------------top 10-----------')
        top = df.head(10)
        print(top)
        # tail 10：获取后10行
        print('------------tail 10------------')
        tail = df.tail(10)
        print(tail)
        # filter：根据指定条件过滤
        print('-------------filter----------')
        special_jobid = df[df.JobID == 595950]
        print(special_jobid)
        special = df[(df.id >= 110) & (df.PublishDate == '2018-12-18')]
        print(special)
        # check ：检查各列缺失情况
        print('-------------check----------------')
        check = df.info()
        print(check)
        print('--------------data describe-----------')
        # count，平均数，标准差，中位数，最小值，最大值，25 % 分位数，75 % 分位数
        describe = df.describe()
        print(describe)
        # 添加新列
        df2 = df.copy()
        df2['AnnualSalaryAvg'] = (df['AnnualSalaryMax'] + df['AnnualSalaryMin']) / 2
        # 重新排列指定列
        columns = ['JobID', 'JobTitle', 'CompanyName', 'AnnualSalaryMin', 'AnnualSalaryMax', 'AnnualSalaryAvg']
        df = pd.DataFrame(df2, columns=columns)
        print(df)
    
    
    """
    可视化分析
    """
    
    
    def visualized_analysis(path):
        df = read_from_csv(path)
        df_post_count = df.groupby('JobLactionStr')['AnnualSalaryMax'].count().to_frame().reset_index()
        # subplots(1, 1) 表示1x1个子图，figsize=(25, 8) 子图的宽度和高度
        # f, [ax1,ax2] = plt.subplots(1, 2, figsize=(25, 8)) 表示1x2个子图
        f, ax2 = plt.subplots(1, 1, figsize=(25, 8))
    
        sns.barplot(x='JobLactionStr', y='AnnualSalaryMax', palette='Greens_d', data=df_post_count, ax=ax2)
        ax2.set_title('各城市职位数量对比', fontsize=15)
        ax2.set_xlabel('城市')
        ax2.set_ylabel('数量')
        # 用来正常显示中文标签
        plt.rcParams['font.sans-serif'] = ['SimHei']
        # 用来正常显示负号
        plt.rcParams['axes.unicode_minus'] = False
        plt.show()
    
    
    if __name__ == '__main__':
        path = 'G:/zhilian.csv'
        read_to_csv(0, 100, path)
        df = read_from_csv(path)
        simple_op(path)
        visualized_analysis(path)

控制台输出如下：

![watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3p3eDE5OTIxMjE1_size_16_color_FFFFFF_t_70][]

![watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3p3eDE5OTIxMjE1_size_16_color_FFFFFF_t_70 1][]

写入本地csv文件效果如下：

![watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3p3eDE5OTIxMjE1_size_16_color_FFFFFF_t_70 2][]

可视化效果如下：

![watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3p3eDE5OTIxMjE1_size_16_color_FFFFFF_t_70 3][]

注：由于是入门文章第一篇，所以并没有对数据分析做过深的探索，更深层次的研究将会在后续系列中呈现！

[watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3p3eDE5OTIxMjE1_size_16_color_FFFFFF_t_70]: /images/20220403/d10ce7b4067047f4996dd5748df22fa8.png
[watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3p3eDE5OTIxMjE1_size_16_color_FFFFFF_t_70 1]: /images/20220403/4b25cee936d94d77894c2900a7d11273.png
[watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3p3eDE5OTIxMjE1_size_16_color_FFFFFF_t_70 2]: /images/20220403/8423745ce2a44f9daecd7fed716c62e3.png
[watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3p3eDE5OTIxMjE1_size_16_color_FFFFFF_t_70 3]: /images/20220403/0f111cb32dd84fbba371c58a99b033ba.png