The Di-Tech Challenge is organised by the Didi Research Institute and open to big data and machine learning experts from around the world. Through accessing real data, participants will have a chance to improve upon a core algorithm in Didi’s product, affecting the experience of hundreds of millions of users across the world.



As less than 10% of China’s 1.4 billion citizens own automobiles, the frequency at which Chinese citizens commute on taxis, buses, trains, and planes is the highest in the world. Didi Chuxing, the dominant ride-hailing company in China, processes over 11 million trips, plans over 9 billion routes and collects over 50TB of data per day. To meet needs of riders, Didi must continually innovate to improve cloud computing and big data technologies and algorithms in order to process this massive amount of data and uphold service reliability.
Supply-demand forecasting is critical to enabling Didi to maximise utilisation of drivers and ensure that riders can always get a car whenever and wherever they may need a ride. Supply-demand forecasting helps to predict the volume of drivers and riders at a certain time period in a specific geographic area. For instance, demand tends to surge in residential areas in the mornings and in business districts in the evenings. Supply-demand forecasting allows Didi to predict demand surges and guide drivers to those areas. The end result is higher earnings for drivers and no surge pricing for riders!
File name (final data set released on 07/05)new ! File format Download
training_set.tar.gz .tar.gz(600MB) Download
test_set.tar.gz .tar.gz Download


Grand Prize
First Round Finalists
Each member of the Top 50 teams will be awarded with:

DiTech Challenge Finalist Certificate from DIDI;
Customized, commemorative T-shirt;

Guaranteed first-round interviews for positions at the DiTech Research Institute;

Members of Top 5 teams will be fast-tracked to final interviews for intern positions.
Outstanding Presentation Award
Cash award of RMB ¥30,000
The Most Potential
Cash award of RMB ¥30,000


He Xiaofei
President of the DIDI Research Institute
He Xiaofei is a worldwide renowned scholar in machine learning and an IAPR Fellow. He serves as a member of the editorial board in multiple top-level international academic journals as well as area chair and program committee member of multiple international conferences.
Ye Jieping
Vice President of the DIDI Research Institute
Ye Jieping is a tenured professor of the University of Michigan and a worldwide leader in machine learning and data mining. He serves as an editorial board member in several top-level international academic journals (such as IEEE TPAMI, DMKD) as well as the program chair and area chair of multiple international conferences
Qingxiong Yang
Head of three-dimensional reconstruction studies at Didi Chuxing
Qingxiong was assistant professor for Department of Computing at City University of Hong Kong and earned a Doctoral Degree in Computing Engineering at UIUC. He published 31 papers on CVPR, ICCV, ECCV, IJCV and PAMI. Five of them have been quoted over 2,100 times, 10 over 100 times, 1 more than 600 times.
Sebastian Thrun
Udacity founder and President , GoogleX founder
Sebastian Thrun is founder and president of Udacity, whose mission is to democratize education. Sebastian was the founder of Google X where he led the development of the Google self-driving car, Google Glass, and other projects. Prior to that, he was professor at Stanford and Carnegie Mellon in artificial intelligence and robotics. Thrun has received several awards including the inaugural Smithsonian American Ingenuity Award for Education in 2012. His work has been highlighted in 50 Best Inventions of 2010 by TIME magazine. Thrun has authored over 380 academic papers and several books, and is member of the National Academy of Engineering.
Peter Norvig
President of Google Research Institute
Peter Norvig is a Director of Research at Google Inc. Previously he was head of Google's core search algorithms group, and of NASA Ames's Computational Sciences Division, making him NASA's senior computer scientist. He received the NASA Exceptional Achievement Award in 2001. He has taught at the University of Southern California and UC Berkeley, from which he received a Ph.D. in 1986. His publications include Artificial Intelligence: A Modern Approach, Paradigms of AI Programming, and Verbmobil: A Translation System for Face-to-Face Dialog. He is a fellow of the AAAI, ACM, California Academy of Science and American Academy of Arts & Sciences.
Katie Malone
Data scientist
After majoring in engineering physics at Ohio State, Katie moved to Geneva, Switzerland to work on Higgs boson searches at the Large Hadron Collider at CERN. She completed a PhD in particle physics in 2015 at Stanford. During her time at Stanford, she took a summer away from physics to work at the online education startup Udacity. While there, she built an online machine learning course and started a data science podcast. Katie continues that work as a Data Scientist at Civis Analytics, where she is part of the team that researches and develops new kinds of data science tools.


·There is no limit on the number of people on any given team. The team name shall not comprise any terms that violate applicable laws and regulations, public social orders, or good social customs.
·Teams are allowed to merge until 10 days before the first round submission deadline. An individual may join a team (1+n), or teams may merge (n+n). The participants are not allowed to join more than one team at the same time.
·Responsibility (rights) of team leader: A team leader may invite other individuals or teams to join them. However, an individual shall not be a member of any other team when being invited. The title of team leader is transferable.


·The contest is open to all individuals.
.Employees (including interns, part-time employees and reviewers) of Didi/Udacity or their immediate family can take part in the contest. Teams with Didi/Udacity employees will be listed on the leaderboard but not eligible for cash rewards. In the case that a Didi/Udacity team ranks in the top three, the cash award will be granted to the next team in the ranking.
1、No one shall register with more than one account. Teams found to have contestants with multiple accounts will be eliminated from the contest.
2、Participants’ identification information must be correct and valid. Discovery of invalid information will results in disqualification and prizes will be revoked.
3、Deadline for registration, change of team members and identity verification is 12:00 PM on June 8, 2016 (UTC+8), 10 days before the deadline of the first round.
4、Registration: Participants may register for the contest using a personal email account after verifying their email address.
5、official contact information: QQ Group:161141063,WeChat Public Account:diditalent,
6、Participants are prohibited from sharing data or code outside of the contest forum without permission from the Contest Organising Committee. Anyone who violates this rule will be disqualified and may be charged for legal liabilities.
Participants who would like to share their code/model may apply to the Organising Committee via email Exceptional models will be shared on the contest website, and the most outstanding model will be awarded a cash prize of RMB 30,000.
7、Team Mergers: The leader of Team A may generate an invitation code and share it with Team B, which will be merged to Team A upon entering the invitation code. The invitation code is one-time use only and valid for a limited period of time.
8、The contest will be based on the results and all submissions:
The leaderboard will be updated in everyday to show the rank of all teams.