How AI is changing the mobile transportation app: The case of Didi
A few weeks ago, Didi entered the Costa Rican market through an aggressive advertising campaign. Dozens of posters covering the capital, prompting us to download the app for free travel immediately.
Having never heard of Didi before, and after some research, we discovered this Chinese giant, a direct competitor of Uber. A Chinese giant led by a woman, Jean Liu.
Jean Liu is a visionary who dreams of seeing decongested cities tomorrow, with fluid and efficient traffic. No more drivers, no more accidents, and no more waiting. You order your DiDi and in a few minutes; it takes you wherever you want.
The use of machine learning, the future of transport, and cities?
The application of machine learning in many aspects of our lives is no longer a mystery. All industries can use these concepts to optimize their resources, facilitate their logistics, or increase their market share.
In the case of Didi, the number of data at their disposal is incredible. Their CTO Bob Zhang explains in an interview: “Didi operates 25m daily rides driven by around four million drivers. Together, these throw up more than 70 terabytes, covering everything from where drivers go to how fast they drive. Poring through all this, Didi’s machine learning algorithms “can show the future,” Zhang says.
The use of all this data enables technological prowess and improved user experience.
The anticipation of travel demand Their decision-making platform can predict with 15 minutes of expectation that a user will request a driver, having just 85% of the time. Bob Zhang hopes that this anticipation will be an hour or even a whole day after having trained enough “Didi Brain”.
Travel comfort Didi is working on developing its cars and developing an overall “safety” aspect. By analyzing acceleration and braking, by collecting data from the various sensitive sensors, Didi hopes to be able to make the journeys of its users as pleasant as possible.
Optimization of journeys Thanks to the numerous maps available in real-time and data concerning the state of transport, Didi will be able to optimize trips to the maximum. Artificial Intelligence will be able to program the most optimum route and minimize the waiting time, both for the driver and the passenger.
While Didi seems to be in an unprecedented effervescence, continually innovating and conquering new markets (Mexico and Costa Rica lately), is it the same for its competitors?
Access to artificial intelligence methods is not only reserved for Didi, and its competitors also use its technologies to improve their offer.
Uber AI is the research center of Uber concerning artificial intelligence, from fundamental research to the optimization of algorithms, this entity covers many fields of study.
One of his latest successes is “One-click Chat.” One-Click Chat is a tool to facilitate communication between the passenger and the driver. We have all happened to wait for our Uber and see it moving in the other direction or very slowly. One-Click Chat allows you to predict the answers to the question sent by the user to prevent the driver from jeopardizing his safety by writing a message.
One of their latest innovations on Lyft is the optimization between supply and demand. Their PPZ Map model allows them to encourage drivers to identify high demand areas by rewarding them with bonuses. All the information collected permits them a better distribution of drivers.