# -*- coding: utf-8 -*-
# author：baoshan

import pandas as pd
import matplotlib.pyplot as plt
plt.rcParams['font.sans-serif'] = ['SimHei']
plt.rcParams['axes.unicode_minus'] = False


def main():
    aqi_data = pd.read_csv('china_city_aqi_teacher.csv')
    print('基本信息:')
    print(aqi_data.info())
    print('数据预览')
    print(aqi_data.head())

    # 数据清洗
    # 只保留AQI＞0的数据
    filter_contation = aqi_data['AQI'] > 0
    clean_data = aqi_data[filter_contation]

    # 基本统计
    print('AQI最大值', clean_data['AQI'].max())
    print('AQI最小值', clean_data['AQI'].min())
    print('AQI均值', clean_data['AQI'].mean())

    # top50
    top50_cities = clean_data.sort_values(by=['AQI'], ascending=True).head(10)
    print('空气质量最好的十个城市')
    top50_cities.plot(kind='line', x='City', y='AQI', title='空气质量最好的50个城市', figsize=(20,10))
    plt.savefig('top50_aqi_line.png')
    top50_cities.plot(kind='bar', x='City', y='AQI', title='空气质量最好的50个城市', figsize=(20, 10))
    plt.savefig('top50_aqi_bar.png')
    plt.show()


if __name__ == '__main__':
    main()

 

pandas内置matplotlib绘图更简单。