Synthetic Control Method

[1]:

import os
import normet
import pandas as pd
import numpy as np
import matplotlib.pyplot as plt

[2]:

conda install openpyxl

Channels:
 - defaults
 - conda-forge
Platform: osx-arm64
Collecting package metadata (repodata.json): done
Solving environment: done

## Package Plan ##

  environment location: /Users/user/anaconda3/envs/normet

  added / updated specs:
    - openpyxl


The following packages will be downloaded:

    package                    |            build
    ---------------------------|-----------------
    et_xmlfile-1.1.0           |  py312hca03da5_1          12 KB
    openpyxl-3.1.2             |  py312h80987f9_0         675 KB
    ------------------------------------------------------------
                                           Total:         687 KB

The following NEW packages will be INSTALLED:

  et_xmlfile         pkgs/main/osx-arm64::et_xmlfile-1.1.0-py312hca03da5_1
  openpyxl           pkgs/main/osx-arm64::openpyxl-3.1.2-py312h80987f9_0

The following packages will be SUPERSEDED by a higher-priority channel:

  certifi            conda-forge/noarch::certifi-2024.6.2-~ --> pkgs/main/osx-arm64::certifi-2024.6.2-py312hca03da5_0



Downloading and Extracting Packages:
openpyxl-3.1.2       | 675 KB    |                                       |   0%
openpyxl-3.1.2       | 675 KB    | 8                                     |   2%
et_xmlfile-1.1.0     | 12 KB     | ##################################### | 100%


Preparing transaction: done
Verifying transaction: done
Executing transaction: done

Note: you may need to restart the kernel to use updated packages.

[3]:

df=pd.read_excel(r'data/AQ_Weekly.xlsx',parse_dates=['date'])

[4]:

df=df.query(f"date>='2015-05-01'").query(f"date<'2016-04-30'")

[5]:

control_pool=["Dongguan", "Zhongshan" , "Foshan", "Beihai"
               , "Nanning","Nanchang" , "Xiamen", "Taizhou"
               , "Ningbo","Guangzhou" , "Huizhou", "Hangzhou"
               , "Liuzhou", "Shantou", "Jiangmen", "Heyuan", "Quanzhou","Haikou" , "Shenzhen", "Wenzhou", "Huzhou"
               , "Zhuhai", "Fuzhou", "Shaoxing", "Zhaoqing","Zhoushan"
               , "Quzhou", "Jinhua", "Shaoguan" , "Sanya"
               , "Jieyang" , "Meizhou", "Shanwei"
               , "Zhanjiang" , "Chaozhou", "Maoming" , "Yangjiang"]

[6]:

df=df[df['ID'].isin(control_pool+["2+26 cities"])]

[7]:

df.head()

[7]:

	date	ID	CO	COwn	NO2	NO2wn	O3	O3_8h	O3_8hwn	O3wn	Ox	Oxwn	PM10	PM10wn	PM2.5	PM2.5wn	SO2	SO2wn
552	2015-05-03	2+26 cities	1.277812	1.328918	38.484623	45.068304	83.690318	133.060175	90.980089	65.825884	60.984249	54.415501	134.523758	139.447383	82.221006	73.876656	32.971551	36.822724
560	2015-05-10	2+26 cities	1.113243	1.298979	35.289414	44.352582	70.075957	103.363985	90.875789	66.100641	52.659234	54.337364	111.771975	134.412595	58.313419	69.386309	27.898696	35.284317
568	2015-05-17	2+26 cities	0.996762	1.274902	35.810778	43.981644	83.598553	125.082867	91.245896	66.418891	59.588514	54.389203	117.075504	129.474025	52.802738	67.146002	29.642007	34.742172
576	2015-05-24	2+26 cities	1.031094	1.267058	39.838150	43.398110	96.838000	149.709652	92.792728	66.695864	68.180477	54.529002	117.650986	121.537391	61.278043	65.876502	34.972942	34.482513
584	2015-05-31	2+26 cities	1.297053	1.285556	37.988483	43.548340	109.770206	165.029315	92.427553	66.419975	73.600167	54.475052	143.115653	126.681810	88.963408	67.239277	33.903061	34.490734

[8]:

df['after_treatment']=df['date']>'2015-10-23'

[9]:

xx=normet.scm(df,'SO2wn','date','ID',"2+26 cities",control_pool,'after_treatment')

[10]:

xy=normet.scm_parallel(df,'SO2wn','date','ID',control_pool,'after_treatment')

[11]:

xx[['SO2wn','synthetic']].plot()

[11]:

<Axes: xlabel='date'>

[12]:

fig,ax=plt.subplots()
for i,city in enumerate(list(xy['ID'].unique())):
    xy[xy['ID']==city]['effects'].plot(ax=ax)

[ ]: