Skip to content

Estimating the effect of Covid-19 Precautions on traffic accidents in Taoyuan City with Regression Discontinuity Design using Python PanelOLS model.

Notifications You must be signed in to change notification settings

jycc-267/Effects-of-COVID-19-Precautions-on-Traffic-Accidents-Taoyuan

Repository files navigation

Effects-of-COVID-19-Precautions-on-Traffic-Accidents-Taoyuan

本專案感興趣的問題是中央政府應對新冠疫情實施二級警戒對於桃園市交通事故發生的外部效果為何?其交通事故肇事原因與駕駛人年齡、性別、是否酒駕等分布是否造成與以往不同之趨勢?在觀察跨期的事故發生資料時,二級警戒實施與否的資料樣本特徵會有明顯的不同,對於應變項的影響也會不一樣。對於這些變數篩選問題,模型採用斷點迴歸設計(Regression Discontinuity)測量二級警戒對於交通事故發生的影響。此外,不方便觀察的控制變數或中介變數,例如每天的車流量、尖峰離峰的時間等,則加入時間固定效果處理(Adding Time Fixed Effect with PanelOLS)。實證結果顯示,疫情警戒確實使交通事故大幅下降、超速案件比重亦確實上升,惟死亡案件比重未有顯著差異。

  • Constructed Regression Discontinuity Design with Time Fixed Effect on public API data from the Taoyuan City government to estimate the causal effect of COVID-19 alert levels on traffic accident outcomes using Python PanelOLS package.
  • Identified a significant decrease in daily traffic accidents, averaging 43 fewer accidents per day, following the implementation of Covid-19 alert levels, while the proportion of speeding cases increased significantly.

Code and Visualization

About

Estimating the effect of Covid-19 Precautions on traffic accidents in Taoyuan City with Regression Discontinuity Design using Python PanelOLS model.

Topics

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published