Dr Cobus Burger, a postdoctoral fellow in the Department of Economics, was named the winner of an international data science challenge aimed at helping the South African government reduce road accident fatalities.
The competition – from October 2019 to early February 2020 – attracted 738 data scientists from across the continent and the world. Its objective was to build a machine learning model that accurately predicts when and where the next road incident will occur in Cape Town, using historic road incident data as well as traffic data from the Uber platform.
The resulting model will enable South African authorities to anticipate where they will be needed next and to put measures in place that will help ensure safety on Cape Town’s roads.
The Transport Economics group within the Department of Logistics ran the competition in partnership with Uber, South African National Roads Agency Limited and data science competition platform Zindi.
Burger, who also works as a senior data scientist at Predictive Insights, said his winning entry was a single light GBM model that focused on the last three months of 2018. (Light GBM is a machine learning algorithm.)
He explained: “Almost all the explanatory power of my model came from the camera data. Historic data were used to approximate how many cars will be using each road segment for each hour of each day of the week. These forecasts were highly correlated with traffic incidents.
“I used the camera names to distinguish between inbound and outbound cameras. I found this distinction to be very useful. I also down-weighted days which had no camera data since these days were informative.”