pm2.5

五城PM2.5分析

① csv文件数据处理

• 名词解释

名词	释义
DEMP	露点
HUMI	湿度
PRES	气压
TEMP	温度
cbwd	风向
Iws	累计风速
Precipitation	降雨量
Ipre	累计降雨量

• 数据导入

  pd.read_csv
  # 我无法使用相对路径，每次都用的绝对路径，很不简洁
  

---

• 粗处理
  • 或许应该采用：

  PM2.5_data_dict={'beijing': ('略', ['PM_Dongsi', 'PM_Dongsihuan', 'PM_Nongzhanguan'], '北京'),
    'chengdu': ('', ['PM_Caotangsi', 'PM_Shahepu'], '成都'),
    'guangzhou': ('', ['PM_City Station', 'PM_5th Middle School'], '广州'),
    'shanghai': ('', ['PM_Jingan', 'PM_Xuhui'], '上海'),
    'shenyang': ('', ['PM_Taiyuanjie', 'PM_Xiaoheyan'], '沈阳')}
    #但我遇到bug：只能读取最后一条键值对
    #所以后面基于dict的处理都只能输出shenyang的数据……
  

  • 我的做法

  #导入后未将数据整合，逻辑简单了但重复的操作增多
  

  #错了又错，最后找到方法了11.23
  

  #修改了season列，显示具体的季节名称而非序号
    df_bj['season'] = df_bj['season'].map(
    {1: 'Spring', 2: 'Summer', 3: 'Autumn', 4: 'Winter'})
    df_sh['season'] = df_sh['season'].map(
    {1: 'Spring', 2: 'Summer', 3: 'Autumn', 4: 'Winter'})
    df_sy['season'] = df_sy['season'].map(
    {1: 'Spring', 2: 'Summer', 3: 'Autumn', 4: 'Winter'})
    df_cd['season'] = df_cd['season'].map(
    {1: 'Spring', 2: 'Summer', 3: 'Autumn', 4: 'Winter'})
    df_gz['season'] = df_gz['season'].map(
    {1: 'Spring', 2: 'Summer', 3: 'Autumn', 4: 'Winter'})
  
  

---

② 计算与分析

• 计算均值
  • 题目要求：计算各城市每日均值

  #一开始采用了这种傻瓜且错误的办法，没有达到预期效果并且使数据更复杂，同时还创建了一个新的csv文件用于存放新列
  df_bj['PM_Meanbj'] = round((df_bj['PM_Dongsi']+df_bj['PM_Dongsihuan']+df_bj['PM_Nongzhanguan'])/3)
    df_cd['PM_Meancd'] = round((df_cd['PM_Caotangsi']+df_cd['PM_Shahepu'])/2)
    df_gz['PM_Meangz'] = round(
        (df_gz['PM_5th Middle School']+df_gz['PM_City Station'])/2)
    df_sh['PM_Meansh'] = round((df_sh['PM_Jingan']+df_sh['PM_Xuhui'])/2)
    df_sy['PM_Meansy'] = round((df_sy['PM_Taiyuanjie']+df_sy['PM_Xiaoheyan'])/2)
  

  #改良
  # 以北京为例
  bj_pm = df_bj.loc[:, ['PM_Dongsi', 'PM_Dongsihuan', 'PM_Nongzhanguan']]
  bj_pm.replace(0, np.nan)
  bj_pm_mean = (bj_pm.mean(axis=0)).mean()
  #结果：所有时间的国内气象站测点平均值90.11966033999813
  #并不是daily的
  
  

  • 再改良->得出结果

  dict = {'beijing': ('',['PM_Dongsi', 'PM_Dongsihuan', 'PM_Nongzhanguan'], '北京'), 
        'shanghai': ('', ['PM_Jingan', 'PM_Xuhui'],'上海'),
        'chengdu':('',['PM_Caotangsi','PM_Shahepu'],'成都'),
        'guangzhou':('',['PM_City Station','PM_5th Middle School'],'广州'),
        'shenyang':('',['PM_Taiyuanjie', 'PM_Xiaoheyan'],'沈阳')}
  
  bj_district = dict['beijing'][1]
  bj_day_pm=df_bj.groupby(by=['year','month','day'])[bj_district].mean().mean(axis=1)
  
  

---

• 数据可视化

  • 基于中国地图显示五城pm2.5均值

    • 开始放飞

    # 将数据可视化——PM2.5地理分布图
    # 激活对象
    geo = Geo()
    # 添加中国地图
    geo.add_schema(maptype='china')
    # 添加数据
    geo.add('PM2.5值', [['北京', bj_pm_mean], ['上海', sh_pm_mean], [
            '成都', cd_pm_mean], ['广州', gz_pm_mean], ['沈阳', sy_pm_mean]])
    geo.set_series_opts(label_opts=opts.LabelOpts(is_show=False))
    geo.set_global_opts(
        visualmap_opts=opts.VisualMapOpts(is_piecewise=True),
        title_opts=opts.TitleOpts(title='北上广成沈五城PM2.5平均值')
    )
    # 绘制成功
    geo.render()
    

    北上广成沈PM2.5

  • 绘制逐年空气质量变化->matplotlib

    • 第一次尝试，只得出了每个测点每年的平均值

    # 以北京为例
    bj_year_pm = df_bj.groupby(df_bj['year']).mean().loc[:, ['PM_Dongsi', 'PM_Dongsihuan', 'PM_Nongzhanguan','PM_US Post']]
    # 未得出三个测点逐年平均值
    # warning
    # The default value of numeric_only in DataFrameGroupBy.mean is deprecated. In a future version, numeric_only will default to False. 
    #  数据中有空值 在mean中添加
    mean(numeric_only=True)
    

    # 所以试图…
    (93.209072+92.496761+90.041620)/3
    # 91.91581766666667
    # 先这么继续吧，说不定后面就知道该怎么写了
    

    • 经过多次尝试，终于找到方法了->显示年均值

    bj_day_pm=df_bj.groupby(by=['year','month','day'])[bj_district].mean().mean(axis=1)  
    

    #再次修改
    bj_year_pm=df_bj.resample('Y')[bj_district].mean().mean(axis=1)
    

    df_bj.set_index(pd.PeriodIndex(year=df_bj['year'], month=df_bj['month'],
                day=df_bj['day'], hour=df_bj['hour'], freq='H'), inplace=True)
    year_x=bj_year_pm.dropna().index
    bj_year_y=list(bj_year_pm)
    plt.figure(figsize=(12,7))
    plt.title('Year_Average_PM2.5')
    plt.xlabel('year')
    plt.ylabel('PM2.5')
    plt.plot(year_x,bj_year_y,label="beijing")
    plt.plot(year_x,cd_year_y,label='chengdu')
    plt.plot(year_x,sh_year_y,label='shanghai')
    plt.plot(year_x,sy_year_y,label='shenyang')
    plt.plot(year_x,gz_year_y,label='guangzhou')
    plt.legend()
    plt.show()
    

    年平均PM2.5