Hate your roommate? Algorithm helps you find a better one in college

Nanjing University in Jiangsu province is using big data to assign dormitories to freshmen in a bid to help early birds and night owls flock together.

The freshmen can fill in a detailed questionnaire on the university’s website, including when they get up and go to bed, how often they change clothes and their willingness to share electricity and Wi-Fi fees.

The answers will then be put into an algorithmic model and the similarities between data sets will be analyzed so that students with similar habits can be assigned to the same dormitories.

Guo Yamin, of the university’s undergraduate admission office, said the questionnaire helps freshmen get along with each other and adapt to university life.

“About 80 percent of the 3,300 freshmen filled out the questionnaire,” Guo said. “We try not to put students with widely different habits into the same dormitory to avoid conflicts and disputes.”

He said the traditional approach alone — putting students with the same listed hobbies together — doesn’t work well because it’s hard to quantify similarities in freshmen’s interests. The new approach adds more criteria.

“The algorithmic model increases our chances of finding potential roommates who can communicate and live harmoniously,” Guo said.

The university revisited the students who were assigned dormitories after completing the questionnaire last year. It found that a larger proportion of them feel satisfied with university life compared with those whose questionnaire results were not applied in roommate selection.

More than 70 percent of university students in China are troubled by conflicts with roommates, according to a survey of 1,355 university students across the country by China Youth Daily in 2015.

It found that the main sources of conflict were different habits, lack of communication and incompatible temperaments.

Gong Yue, director of the university’s student affairs office, said freshmen can learn communication skills and how to interact with others by living with roommates.

“Guidance provided by the university can help them adapt to their new phase of life. It can effectively help the roommates influence learn from each other. ”

The university said that the questionnaire will only be used for assigning dormitories, not for profiling students in other ways.

Guo said that despite the development of the algorithmic model, students should also learn how to live in a group and get along with different people. and Guo Jun contributed to this story.