Geographically weighted logistic regression analysis on achieving air quality index performance targets in East Java

Air quality in an area can be assessed through an air quality index that considers five main types of pollutants, namely nitrogen dioxide, sulfur dioxide, ozone, particulate matter, and carbon monoxide. Indonesia, including East Java Province, experiences significant air pollution problems, according to the Air Quality Life Index (AQLI) Report, which contributes greatly to global air pollution. To achieve a clean, healthy, and harmonious life, the air quality index problem must be addressed immediately, in accordance with the sustainable development goals (SGDs) which include the 3rd, 7th, and 11th SDGs. In this study, Geographically Weighted Logistic Regression (GWLR) analysis is used to analyze the achievement of the air quality index target in East Java Province in 2023. The response variable is the achievement of the air quality index target with a nominal scale of 0 to 1, while the predictor variables include the number of tourism accommodations, population density, forest and water area, and the number of motorized vehicles. The GWLR model with the Fixed Gaussian Kernel function has better performance than the global binary logistic regression model because of the higher accuracy measure of the GWLR model, which is 78.95% in classifying the status of achieving the air quality index target in East Java. These findings provide a spatially-informed approach to policymaking for improving air quality in East Java.