Uncovering Spatial Heterogeneity in Poverty Determinants for Data-Driven Social Welfare Planning

Abstract

Poverty is one of the most profound social problems in Indonesia, with uneven distribution across regions. Some regions in Indonesia, especially in remote areas, still face high poverty rates despite various poverty alleviation efforts. This inequality indicates that socio-economic and geographic factors play a significant role in the distribution of poverty. Therefore, this research aims to analyze the spatial distribution of the percentage of poor people in regencies/ cities in Java in 2025. This research will use secondary data from the Central Bureau of Statistics and spatial analysis using Geographically Weighted Regression (GWR). Through this spatial approach, it is hoped that a deeper understanding of how these factors contribute to poverty can be obtained and whether there are certain patterns that can be identified in each region. The results of the research show that the percentage of poor population is modeled using GWR with a Gaussian kernel weighting function and a fixed bandwidth yields distinct models for each regency/city in East Java for the year 2025. The Maximum Likelihood method is employed in this research to estimate the parameters. The findings of this research will significantly contribute to the development of more targeted policies for poverty alleviation, focusing on interventions tailored to the characteristics of each region. Policy recommendations based on spatial analysis will be more effective in reducing poverty in the most affected areas and improving access to public services and infrastructure in underdeveloped regions.