Zindi, AWS, SANSA Develop Model to Identify Informal Settlements

A collaboration between Zindi, Amazon Web Services (AWS) and the South African National Space Agency (SANSA) has yielded a machine learning model to identify informal settlements from satellite imagery. This tool will help the South African government with planning, providing essential services, and preventing crime in underserved communities across the country.

In June 2020, Zindi hosted a hackathon using SANSA’s data and AWS’s virtual machines. The competition platform invited more than 150 data scientists with proven computer vision expertise to participate in the hackathon. They thus competed for USD 1000 and a chance to make the world a better place, using AWS sponsored virtual machines for the computing power needed for machine learning engineering.

Consequently, data scientists from 34 countries used Satellite Pour l’Observation de la Terre “Satellite for observation of Earth” (SPOT) satellite image data provided by SANSA to create the models. Participants used satellite images with manually labelled informal settlements around Johannesburg in Gauteng as training data for modelling. The hackathon challenged data scientists to develop models that can find informal settlements in KwaZulu-Natal. After 60 hours of competition, Raphael Kiminya from Kenya came out on top with a model that managed to predict informal settlements that human labellers had missed. As a result, Zindi and SANSA are working together to put these models into practice.

According to Zindi CEO Celina Lee. “The models that we have created will provide a kind of heatmap, with different probabilities indicating where an informal settlement is likely to be.” She added that “What’s nice is that the model can pick up on what the human eye might just scan right over and not notice.” She also mentioned that it was exciting to work with SANSA on the project as it unlocked opportunities for many African data scientists to showcase their talents.

The Managing Director of Earth Observations at SANSA, Andiswa Mlisa, commented that “The idea was to get data scientists to work on a potential solution that SANSA could use to optimise our mapping processes.” Furthermore, he added that the model would assist SANSA in mapping informal settlements, which SANSA currently undertakes manually.

Zindi is a competition platform hosting a community of data scientists dedicated to solving Africa’s most demanding challenges through machine learning & AI. This article was republished from Medium’s publication. You can read the original article here.


Please enter your comment!
Please enter your name here

This site uses Akismet to reduce spam. Learn how your comment data is processed.