Mapping the World's Offline Population
This project develops a predictive model to estimate the number of people connected to the internet (internet, in this instance, is a proxy for access to social and economic opportunities). The model was trained on a combination of survey data and multiple open source datasets. With the help of the custom engineered data gathering pipeline, the open source data can be collected for any given country and either the existing model can be applied or a new model can be trained if survey data is available.
This project was built from June - August 2021 and is a collaboration between the Data Science for Social Good Fellowship at the University of Warwick, UK and the International Telecommunications Union, a UN Agency.
Click on the pages on the right to learn more about our project and find out how to replicate it for yourself.
Need help getting started?
If you are a policy maker, we recommend you start by reading our introduction and conclusion.
If you are a technical person or otherwise seeking to replicate what we did for your own project, we recommend starting at the Data Gathering section.
If you are curious about who we are or what datasets we've used, start at About the Team or Exploratory Data Analysis.
Watch below to see a five minute presentation about our project: