Using Social Media Big Data and ChatGPT for Identifying Counter-urbanisation Hot Spots in China

Abstract

While urban areas remain home to most of the global population, the countryside is increasingly becoming a preferred place of residence, even in some developing nations such as China. Potential benefits and challenges for rural development posed by this counter-urbanisation trend make it essential to monitor its extend and progression. However, in China, statistical data on this phenomenon remains scarce due to the restrictions on rural property ownership by outsiders, who mostly rent properties from individual rural residents, leaving no formal record. This research uses the data from the Chinese social media platform, XiaoHongShu (or RedNote, a platform similar to Instagram), where specific users share their experience renting houses in rural areas or moving back to home villages. A web crawler was developed to collect all these ‘big data’, which was subsequently cleaned and processed by ChatGPT AI model. To ensure credibility, partial human validation was also applied to AI-generated results. Despite certain limitations, the findings provide original insights into the counterurbanisation trend in China, highlighting its widespread occurrence and identifying several hot spot areas. This study demonstrates that combining social media big data with AI processing offers an effective and timely way to identify human migration flows between urban and rural areas, enabling valuable reference for policymakers and the real estate sector.