TY - JOUR
T1 - Air Quality Modeling Using Real-time Urban Big Data
AU - Jo, Joonsik
AU - Kim, Gahyun
AU - Kwak, Juhyeon
AU - Jeong, Ilho
AU - Ku, Donggyun
AU - Lee, Seungjae
N1 - Publisher Copyright:
Copyright © 2023, AIDIC Servizi S.r.l.
PY - 2023
Y1 - 2023
N2 - This study utilized real-time urban data from Seoul to assess the impact of transportation mode choice on air pollution metrics. EXtreme Gradient Boosting (XGBoost), an ensemble model, was employed for air quality analysis. Subsequently, SHapley Additive exPlanations (SHAP), one of eXplainable Artificial Intelligence (XAI) techniques, was used to understand the influence of urban factors on air quality. For spatial coverage, 50 locations with high traffic volume were selected, and the temporal coverage spanned from April 3, 2023, to April 30, 2023. Variables related to traffic, the environment, and weather were established as features, while Comprehensive Air-quality Index (CAI), PM2.5, and PM10 were determined as target variables. As a result, Root Mean Squared Error (RMSE) of the models predicting CAI, PM2.5, and PM10 were calculated as 0.57, 0.47, and 0.50. The study found that as the maximum number of pedestrians and the number of subway passengers alighting increased, the levels of CAI, PM2.5, and PM10 decreased. This indicates that the use of greener modes of transportation, such as walking and taking the subway, positively impacts air pollution reduction In addition, lower road traffic speeds were associated with higher PM2.5 levels, while increased road congestion correlated with higher PM10 levels. The observed increase in PM2.5 and PM10 levels in relation to the rise in passenger car traffic suggests that emissions from these vehicles are contributing to air pollution. Consequently, the study confirmed that traffic-related factors can influence air quality indicators, and that modifications to traffic volumes and modal splits can enhance air pollution control. This study provides a foundation for developing policies to improve air quality by quantifying and presenting various factors that impact air quality.
AB - This study utilized real-time urban data from Seoul to assess the impact of transportation mode choice on air pollution metrics. EXtreme Gradient Boosting (XGBoost), an ensemble model, was employed for air quality analysis. Subsequently, SHapley Additive exPlanations (SHAP), one of eXplainable Artificial Intelligence (XAI) techniques, was used to understand the influence of urban factors on air quality. For spatial coverage, 50 locations with high traffic volume were selected, and the temporal coverage spanned from April 3, 2023, to April 30, 2023. Variables related to traffic, the environment, and weather were established as features, while Comprehensive Air-quality Index (CAI), PM2.5, and PM10 were determined as target variables. As a result, Root Mean Squared Error (RMSE) of the models predicting CAI, PM2.5, and PM10 were calculated as 0.57, 0.47, and 0.50. The study found that as the maximum number of pedestrians and the number of subway passengers alighting increased, the levels of CAI, PM2.5, and PM10 decreased. This indicates that the use of greener modes of transportation, such as walking and taking the subway, positively impacts air pollution reduction In addition, lower road traffic speeds were associated with higher PM2.5 levels, while increased road congestion correlated with higher PM10 levels. The observed increase in PM2.5 and PM10 levels in relation to the rise in passenger car traffic suggests that emissions from these vehicles are contributing to air pollution. Consequently, the study confirmed that traffic-related factors can influence air quality indicators, and that modifications to traffic volumes and modal splits can enhance air pollution control. This study provides a foundation for developing policies to improve air quality by quantifying and presenting various factors that impact air quality.
UR - http://www.scopus.com/inward/record.url?scp=85183508750&partnerID=8YFLogxK
U2 - 10.3303/CET23106039
DO - 10.3303/CET23106039
M3 - Article
AN - SCOPUS:85183508750
SN - 2283-9216
VL - 106
SP - 229
EP - 234
JO - Chemical Engineering Transactions
JF - Chemical Engineering Transactions
ER -