The rise of self-driving cars and the shift toward shared mobility services is set to dramatically alter the way we move around cities in the future. This should help to reduce traffic in urban areas and cut greenhouse gas emissions. For these services to grow, accurate and computationally efficient algorithms are needed to effectively match individuals with on-demand vehicles, in order to cope with the hundreds of thousands of trips that are routinely made within large cities. Researchers at the Senseable City Lab have now developed an algorithm that represents the shareability of the taxi fleet as a graph. Using this graph, the algorithm was able to find the best solution for fleet sharing. The Minimum Fleet Network model developed by the Senseable City Lab could reduce the taxi fleet size by 40%.