Productive Zakat and Data-Driven Strategies as a Solution to Regional Poverty in Indonesia

Abstract

This study investigates the role of zakat in alleviating poverty among asnaf (zakat recipients) by employing a data-driven approach tailored to regional disparities in Indonesia. The research analyzed data collected from 33 provinces and applied a decision tree regression (DTR) model to identify key factors influencing poverty levels. The findings indicate that micro-enterprises exert the most significant impact on poverty reduction, followed by agriculture and Qardhul Hasan. The DTR model demonstrated strong overall performance, explaining approximately 97.3% of the variance in the dataset, as indicated by the R² value. However, it exhibited limitations in predicting instances with very low values, reflected by a high mean absolute percentage error (MAPE) of 301.22%. This discrepancy suggests potential challenges when evaluating model accuracy for near-zero target values. These results underscore the importance of contextualizing zakat distribution strategies to address local socioeconomic conditions, particularly through targeted support for micro-enterprises, education, and agricultural development. The study recommends that zakat institutions and local governments utilize empirical data to design more effective and evidence-based policies aimed at poverty alleviation. By emphasizing productive zakat programs, this research contributes practical insights to enhancing the efficiency and sustainability of zakat-based social welfare initiatives.