Replication package for Electric vehicle usage reduces urban air pollution: Insights from multi-year nationwide charging records in China

Replication materials for “Electric vehicle usage reduces urban air pollution: Insights from multi-year nationwide charging records in China”. Please download and unzip the file ”Replication_AQ_EV.zip“. Refer to the README for a detailed description of the sample data included in this repository. README --- Organization of repository ---Scripts: scripts for computation, analysis, and visualizationData: input data used in the analysisMap: materials for map creation --- Instructions ---Data processing and analysis are conducted using Stata 18.0 and R 4.2.0. Package dependencies are declared at the top of each script. --- Scripts -----Table1.doUsed to perform the regressions to estimate the impacts of EV charging usage on air quality. This script can be used to estimate the impacts of EV charging amounts on the concentrations of NO2, PM2.5 and O3 with three sets of fixed effects respectively. --Table2.doUsed to perform the regressions to estimate air quality impacts at different granularity levels. This script can perform three sets of regressions: 1) estimating the impacts of the entry of EV charging station; 2) estimating the impacts of the number of EV charging stations; 3) estimating the impacts of the EV charging amounts. --Figure1.RUsed to generate geographic visualizations of EV charging usage and air quality across 334 Chinese cities from January 2016 to December 2023. This script produces four panels: 1) the number of EV charging station installations per city; 2) the average EV charging amount per station; 3) the average NO2 concentration; 4)the average PM2.5 concentration.Each panel includes a map and a subplot displaying trends over time in southern and northern China, divided by the Huai River-Qinling Mountains line. --Figure2.RUsed to produce visualizations of temperature-dependent effects. This script generates three panels: 1) the linear and cubic relationships between temperature and EV charging usage per station; 2) the heterogeneity of the estimated effects of EV charging usage on NO2 across temperature bins; 3) the heterogeneity of the estimated effects on PM2.5 across temperature levels. --Figure3.RUsed to generate visualizations of air quality improvements per unit of EV deployment. This script produces: 1) maps of NO2 and PM2.5 reductions per thousand EV charging stations; 2) maps of NO2 and PM2.5 reductions per 100 MWh of EV charging amounts; 3) bar plots comparing average effects between northern and southern cities; 4) counterfactual estimates of NO2 and PM2.5 concentrations in the absence of EV deployment --- Data ---The dataset used to support the above analyses and visualizations. Specific data usage corresponds to the data-loading modules in the Scripts lines. --- Map ---Materials being used to support the generation of maps.《电动汽车使用降低城市空气污染：基于中国多年全国充电记录的洞察》的复现材料。 请下载并解压文件"Replication_AQ_EV.zip"。请参阅README文件，了解本仓库中包含的示例数据的详细说明。 README --- 仓库组织结构 --- Scripts：用于计算、分析与可视化的脚本文件 Data：分析过程中使用的输入数据集 Map：用于绘制地图的素材文件 --- 使用说明 --- 本项目的数据处理与分析基于Stata 18.0与R 4.2.0完成，各脚本的顶部已声明所需的软件包依赖。 --- 脚本文件 --- --Table1.do 用于执行回归分析，以评估电动汽车（Electric Vehicle, EV）充电使用对空气质量的影响。本脚本可分别通过三组固定效应模型，估算电动汽车充电量对二氧化氮（NO₂）、细颗粒物（PM₂.5）及臭氧（O₃）浓度的影响。 --Table2.do 用于执行回归分析，以评估不同粒度下的空气质量影响。本脚本可完成三组回归任务：1）评估电动汽车充电站投入运营的影响；2）评估电动汽车充电站数量的影响；3）评估电动汽车充电量的影响。 --Figure1.R 用于生成2016年1月至2023年12月期间中国334个城市的电动汽车充电使用情况与空气质量的地理可视化结果。本脚本可生成四个子图面板：1）各城市的电动汽车充电站安装数量；2）单座充电站的平均充电量；3）平均二氧化氮浓度；4）平均细颗粒物浓度。每个面板均包含一张地图，以及以秦岭-淮河线为界的中国南、北方随时间变化的趋势子图。 --Figure2.R 用于生成与温度相关的影响效应可视化结果。本脚本可生成三个子图面板：1）单座充电站的电动汽车充电量与温度之间的线性及三次关系；2）电动汽车充电使用对二氧化氮浓度影响的估计效应随温度区间的异质性；3）电动汽车充电使用对细颗粒物浓度影响的估计效应随温度水平的异质性。 --Figure3.R 用于生成每单位电动汽车部署量带来的空气质量改善的可视化结果。本脚本可生成：1）每千座电动汽车充电站对应的二氧化氮与细颗粒物浓度降幅地图；2）每100兆瓦时（MWh）电动汽车充电量对应的二氧化氮与细颗粒物浓度降幅地图；3）对比中国南、北方城市平均效应的柱状图；4）无电动汽车部署场景下的二氧化氮与细颗粒物浓度反事实估计值。 --- 数据集说明 --- 本数据集用于支撑上述分析与可视化任务，具体的数据使用方式可对应至各脚本中的数据加载模块。 --- 地图素材 --- 用于支撑地图绘制的相关素材文件。