
Thinking Machines have built a tool to help NGOs to more easily link their own internal datasets with external human development datasets. There is a lack of up-to-date geospatial socioeconomic data for both public policy makers to monitor human development indicators and private businesses to plan expansion and investment. OnTrackPH is a software framework that leverages Natural Language Processing (NLP) techniques to accurately match huge number of records across data silos. In 2016, Thinking Machines was able to link three different Philippine government databases and accurately match records in a fraction of the man-hours needed to do so manually. To realize this goal, they are working on two key products: 1) Link Sight: A web application that NGOs can use to easily standardize inconsistently-spelled names of administrative territories (villages, towns, and provinces) in their datasets, so that these can be linked with external datasets. 2) A collated database of village-level human development indicators from several key sources: government-conducted surveys, open-source maps, and AI-generated maps based on satellite imagery. This will enhance the more traditional sources of human development data that are often incomplete or not up to date.