Case Study: Machine learning for classifying road conditions in Tanzania

See the story of the pilot and our insights on AI and machine learning here

“Normally when people apply AI it is to finely crafted data sets. But in the real world of developmental projects, where data has been either opportunely or coarsely collected, lack of data quality becomes crucial. Dealing with this, and designing for biased, sparse, noisy and incomplete data, simply has to become a core part of the workflow in developing contexts.”

At the Frontier Technologies Hub, we work with UKAid to explore how emerging technologies can help to solve the variety of challenges found in developing country contexts.

In this case study, we present the story and reflect on a completed pilot project in Zanzibar, Tanzania, which looked to see if AI and machine learning could be used to automatically assess road quality conditions.

You can find the publication here.