Ghost Cities of China explores the question of Chinese resdiential underutilization through social media data. The project highlights the power of opensource data in discovering areas of risk in China.
Identifying these areas can have an impact on Chinese real estate investors, planners, and individuals who plan to make home purchases in the future.
We collected data for 22 different second and third tier cities across China in order to create a model that is applicable for most major cities. We collected information from social media sites such as Dianping, Weibo, RenRen, and Baidu.
From Dianping data, which is similar to Yelp, we created an 'amenities score' that calculates score for residential areas in a city based on their accessibility to standards amenities for daily life such as schools, hospitals, restaurants, and grocery stores.
After we created the model, we went to China to verify our results by visting our Ghost City sites that resulted from the model. We also talked to different stakeholders such as planners, academics, and developers to understand the conditions from each party's perspective.