We're launching the
AI4D Research Bank in 2023
Thinking Machines Data Science is working with the UNICEF Venture Fund to build the Artificial Intelligence for Development (AI4D) Research Bank.
Thinking Machines
x
Unicef Venture Fund
We're launching the
AI4D Research Bank
in 2023
Thinking Machines Data Science is working with the UNICEF Venture Fund to build the Artificial Intelligence for Development (AI4D) Research Bank.
Learn more
Our goal
Accelerating machine learning adoption in the development sector
The AI4D Research Bank will support data scientists working at the intersection of machine learning (ML) and development.
We hope that access to our code, documentation, and pre-processed datasets will facilitate further research over the long term, and support development sector agencies to make data-driven decisions through ML and open data.
These resources are now open-sourced!
GeoWrangler, a python library for geospatial data analysis
Air quality estimation and poverty mapping models
Pre-processed datasets used in our ML research
What we've worked on
AI4D Research Bank Components
Air Quality Exploratory Research
We used machine learning to estimate haze or particulate matter (PM)2.5 in Thailand. Our team built on Gupta et al’s (2021) research by training a model on satellite-derived data. This includes aerosol information, meteorological factors, and vegetation index to estimate PM2.5 across the country.
Try our training notebook and review our resources!
Geowrangler
GeoWrangler is a pure python library that executes data transformation functions with fewer lines of code. It supports downstream activities like feature engineering and machine learning in geospatial data workflows.
If you’re working with geospatial datasets, try out GeoWrangler!
View in Github
Poverty Mapping
We’re expanding our poverty estimation work in the Philippines to other countries in Southeast Asia. We’ll use the Demographic Health Survey (DHS) and satellite-derived datasets for model training and evaluation
Coming Soon!