Small Area Estimation Approaches Using Satellite Imageries Auxiliary Data for Estimating Per Capita Expenditure in West Java, Indonesia

Introduction/Main Objectives: The economy of a country can determine the welfare of its people. One of the economic indicators in Indonesia is per capita expenditure, which has the lowest estimation at the district level. Background Problems: Sub-district level estimates provide detailed information on inequality that cannot be explained at the district level. Unfortunately, subdistrict level estimates of per capita expenditure in Indonesia have poor Relative Standard Error (RSE) values. Research Method: The Small Area Estimation (SAE) method can improve estimator accuracy on small samples by using auxiliary variable information. Novelty: The existence of big geospatial data such as remote sensing provides an advantage in the efficient use of auxiliary variables. Finding Result: The Empirical Best Linear Unbiased Prediction (EBLUP) model using Nighttime Light Intensity (NTL) as an auxiliary variable provides the best results of the five proposed models. Remote sensing data can potentially be used in SAE auxiliary variables. It provides opportunities for cheaper, faster, and more efficient data collection compared to conventional data.