Template-Type: ReDIF-Paper 1.0
Author-Name: Abbate Nicolás Francisco
Author-Name-First: Nicolás Francisco
Author-Name-Last: Abbate
Author-Name: Gasparini Leonardo
Author-Name-First: Leonardo
Author-Name-Last: Gasparini
Author-Name: Ronchetti Franco
Author-Name-First: Franco
Author-Name-Last: Ronchetti
Author-Name: Quiroga Facundo
Author-Name-First: Facundo
Author-Name-Last: Quiroga
Title: High-Resolution Income Estimates Using Satellite Imagery: A Deep Learning Approach applied in Buenos Aires
Abstract: In this study, we examine the potential of using high-resolution satellite imagery and machine learning techniques to create income maps with a high level of geographic detail. We trained a convolutional neural network with satellite images from the Metropolitan Area of Buenos Aires (Argentina) and 2010 census data to estimate per capita income at a 50x50 meter resolution for 2013, 2018, and 2022. This outperformed the resolution and frequency of available census information. Based on the EfficientnetV2 architecture, the model achieved high accuracy in predicting household incomes ($R^2=0.878$), surpassing the spatial resolution and model performance of other methods used in the existing literature. This approach presents new opportunities for the generation of highly disaggregated data, enabling the assessment of public policies at a local scale, providing tools for better targeting of social programs, and reducing the information gap in areas where data is not collected.
Length:  19 pages
Creation-Date:  2024-11
File-URL: https://aaep.org.ar/works/works2024/4701.pdf
File-Format: Application/pdf
Number: 4701
Classification-JEL: C81, C45
Handle: RePEc:aep:anales:4701