Mapping Poverty through Data Integration and Artificial Intelligence: A Special Supplement of the Key Indicators for Asia and the Pacific
SHARE THIS PAGE
This special supplement to the Key Indicators for Asia and the Pacific 2020 discusses how poverty estimates can be enhanced by integrating household surveys and censuses with data extracted from satellite imagery.
As part of a special ADB knowledge initiative, computer vision techniques and machine-learning algorithms were applied on datasets from the Philippines and Thailand to demonstrate increased granularity of poverty estimation using artificial intelligence. The report identifies practical considerations and technical requirements for this novel approach to mapping the spatial distribution of poverty. It also outlines the investments required by national statistics offices to fully capitalize on the benefits of incorporating innovative data sources into conventional work programs.
- Estimating Poverty Using Conventional Data Sources
- Using Big Data to Enhance Development Statistics
- Predicting Poverty Using Geospatial Data
- Using Neural Networks to Develop an Algorithm
- Understanding Convolutional Neural Networks
- Outlining Data Requirements for the Feasibility Study
- Applying a Convolutional Neural Network to Poverty Prediction
- Extracting Features from the Convolutional Neural Network
- Predicting Poverty from Features Using Ridge Regression Models
- Outlining the Key Findings of the Feasibility Study
- Preparing National Statistics Offices for the Use of Big Data
- Summary and Conclusion