In the past five and a half years, Didi Chuxing has served more than 450 million users with a full range of mobility services, including Taxi, Express, Premier, Luxe, Hitch, Bus, Minibus, Designated Driving, Car Rental, Enterprise Solutions and Bike-Sharing.

DiDi announced the global expansion of the GAIA Initiative at the 97^th Annual Meeting of The Transportation Research Board (TRB 2018) in Washington, DC. The GAIA Initiative was first launched in China in October 2017, under which scientists may apply for access to the anonymized data to explore solutions to traffic challenges.

Initially, GAIA will provide a dataset containing anonymized start points, end points and routing information of trips on DiDi Express and DiDi Premier, two of DiDi’s core personal mobility services.

For experts in the field of AI and transportation infrastructure, large amounts of high-quality data are critical for research and development of new products and systems. By leveraging DiDi’s big data and the expertise of researchers worldwide, DiDi hopes to raise public awareness of sustainable transportation planning and stimulate the creation of new areas of research and innovation. Any data made available will be anonymized and aggregated, as well as encrypted at rest and in transit.

Prior to the GAIA Initiative, DiDi launched a data visualization platform for city transportation managers to facilitate research of transportation needs and the development of new traffic management strategies. GAIA extended DiDi’s outreach to the academic community within China, and now worldwide, to catalyze broader, deeper research in an expanding range of topics in the field of transportation.

DiDi itself uses the large amounts of data and its intelligent algorithms in a number of ways:

• Estimated time of arrival. Based on massive amounts of real-time travel data, DiDi designs new and significantly more accurate time estimation algorithms which overcomes the shortcomings of traditional computing methods.

• Route planning. Leveraging a large volume of data, DiDi designs innovative route planning algorithms which better simulate decision-making processes of experienced drivers and bring users most efficient mobility options while enhancing overall traffic effectiveness.

• Supply and demand forecasting. Based on a huge volume of trip requests and driver location data, DiDi develops algorithms to forecast the demand of passengers and supply of driver-vehicles in different areas at any time to provide the most efficient mobility options.

• Transportation capacity management. Based on supply-and-demand forecasts, the algorithmic systems orchestrate the full range of available vehicle and driver capacities to optimize resource allocation within the entire city. Real-time and predictive dispatching planning enables us to solve ongoing or even potential supply-demand imbalances, so as to improve efficiency on the platform while optimize resource use while alleviating urban traffic congestion.

• Traffic congestion. Through big data analysis, DiDi can accurately assess traffic congestion in terms of time and location (for example, a day, time period or region, etc.) and provide useful information for city planning and administration.

For planners, the large data set offers insights into:

• Smart traffic lights planning. Traffic lights will be optimized based on the intelligent traffic cloud, route start and end points from DiDi, as well as traditional traffic data sources from the government. By predicting traffic patterns, smart traffic lights will properly manage road space and traffic speed based on traffic flow in the region.

• Public transport capacity management. Analysis of the data can suggest optimized bus routes and departure intervals by analyzing supply and demand of transportation services, facilitating work efficiency improvements for both public and private transport companies.

• Reversible lanes. DiDi can identify streets that could benefit from reversible lane installation. With precise data analysis, DiDi can locate accurate start and end points for reversible lanes and suggest the exact time periods for their operation.

• Safety assessment. Big data-driven deep learning analytics are powerful tools in detecting and identifying driving risk factors. This makes it possible for us to engage in precision intervention upon early warnings and attain higher safety standards with AI capabilities.