Researchers collected data for 0.63 million restaurants between 2017 and 2018 from a restaurant rating website in China on cuisine type, price, taste, and number of reviews, and combined the data with mobile phone location data for 56.3 million people, bank card records, and company registration records from 9 Chinese cities to predict population size, number of firms, and consumption volume; using the method, the authors predicted daytime and nighttime populations, which are proxies for employment and residents, respectively, and the findings also suggest that models trained with restaurant data in different cities can achieve population prediction power across cities, according to the authors.
###
Article #19-03064: "Predicting neighborhoods' socioeconomic attributes using restaurant data," by Lei Dong, Carlo Ratti, and Siqi Zheng.
MEDIA CONTACT: Siqi Zheng, Massachusetts Institute of Technology, Cambridge, MA; tel: 617-806-6756; email: sqzheng@mit.edu
Journal
Proceedings of the National Academy of Sciences