First launched in August 2017 in Johannesburg and Tshwane, Movement, a free public website, addresses specific problems faced by urban planners, city staff, and the civic community.
Uber Movement shows travel times across zones in a city derived from data from the billions of rides that riders have taken with Uber, anonymised and aggregated into zones covering the standard boundaries used by urban planners.
The privacy of both riders and driver-partners has been upheld by ensuring that the data is anonymous and aggregated and cannot be used to access any personally identifiable information or user behaviour. The introduction of Movement data is important for community members, city and transport planners as it will provide them with a more improved data-driven transportation policy that will effectively show them where investments in transportation infrastructure should be made in their cities.
Head of public policy Uber South Africa, Yolisa Kani says: "The launch of Movement has successfully demonstrated Uber’s continued support and commitment to its cities and communities. This tool was designed to effectively evaluate where cities can make investments in transportation infrastructure using real-time data. The Machine Learning hackathon allowed data scientists to put this tool to the test and provided them with a realistic demonstration of how the Movement data tool will work. It has been amazing to witness what this tool will do for the city of Cape Town. With data at our fingertips, we hope to see that Cape Town will have a more reliable and efficient mode of transportation."
With the support of Zindi, Uber recently hosted a one-day Machine Learning Hackathon at the Stellenbosch LaunchLab. The hackathon saw young data scientists in South Africa, come together in teams to build machine learning models using SANRAL data from the traffic command centre and Uber movement data to solve traffic-related issues in the Western Cape.
At the Movement launch event, Zindi CEO, Celina Lee led, Data Saiyans, one of the teams that took top honours at the Machine Learning Hackathon, on presenting their winning strategy on how they used the data provided to them to best predict the probability of when an incident would occur.
From the pool of data provided the young scientists used factors such as weather conditions, structures of the roads, occurrence and frequency of accidents to best solve their challenge.
Siphesihle Yapi, member of one of the winning hackathon teams adds: "Through these innovative solutions we are able to improve the townships in which we reside in to help them become smarter and ultimately enhance road safety."
