Predicting socioeconomic indicators using transfer learning on imagery data: an application in Brazil.

Area of Science:

• Remote Sensing
• Geospatial Analysis
• Socioeconomic Data Analysis

Background:

• Traditional socioeconomic data collection via censuses is costly and time-consuming, leading to significant data gaps in developing countries.
• These data gaps impede effective public policy formulation and research development.
• Satellite imagery presents a potential solution for estimating socioeconomic variables, offering accessible and free data, though requiring remote sensing knowledge.

Purpose of the Study:

• To estimate average income, GDP per capita, and a water index at the city level in Brazil using satellite imagery.
• To apply transfer learning techniques to daytime and nighttime satellite data for socioeconomic variable estimation.
• To assess the efficacy of satellite-derived socioeconomic indicators in two Brazilian states: Bahia and Rio Grande do Sul.

Main Methods:

• Utilized daytime and nighttime satellite imagery.
• Applied a transfer learning approach to analyze the imagery.
• Focused on city-level estimations for socioeconomic indicators in Bahia and Rio Grande do Sul, Brazil.

Main Results:

• The transfer learning model explained up to 64% of the variation in city-level socioeconomic variables.
• Results demonstrated consistency with existing literature, despite potential cross-country data variations.
• This study represents a pioneering analysis of its kind for Brazil.

Conclusions:

• Satellite imagery, combined with transfer learning, provides a viable method for estimating key socioeconomic indicators in Brazil.
• This approach can help mitigate the challenges posed by infrequent and expensive traditional data collection methods.
• The findings are encouraging for future research and policy applications in Brazil and potentially other developing nations.