A Guidebook on Mapping Poverty through Data Integration and Artificial Intelligence
This guidebook identifies tools and resources that can help generate poverty statistics using satellite imagery, geospatial data, and machine-learning algorithms to augment conventional data collection and sample survey techniques.
The “leave no one behind” principle of the 2030 Agenda for Sustainable Development requires appropriate indicators to be estimated for different segments of a country’s population. The guidebook was based on a feasibility study by ADB, in collaboration with the Philippine Statistics Authority, the National Statistical Office of Thailand, and the World Data Lab, that aimed to enhance the granularity, cost-effectiveness, and compilation of high-quality poverty statistics. It also serves as an accompanying guide to the Key Indicators for Asia and the Pacific 2020 special supplement focusing on mapping poverty estimates.
- Hardware and Software Requirements and Setup
- Data Preparation
- Training of Convolutional Neural Network
- Convolutional Neural Network Model Feature Extraction
- Ridge Regression
- Rescaling of Poverty Estimates and Visualization