Are people happier in locations of high property value? Spatial temporal analytics of activity frequency, public sentiment and housing price using twitter data

Abstract

The rise of social networking platforms provides opportunities to examine the relationship between public emotion and housing price. This study investigates frequency and places of visits, population sentiment, and housing price using 8.7 million tweets retrieved from Manhattan, New York City in 2019. We implemented kernel density estimation, Getis-Ord Gi hot spot analysis, and spatial lagged hedonic pricing models to identify the location variation of sentiment levels. The results show: (1) the spatial clustering of tweets frequency was highly related to land use types in places such as parks, financial districts, and train stations; (2) high sentiment levels coincided with high frequency clusters and higher positive sentiment is associated with higher housing price; and (3) sentiment level was significantly associated with housing price and building structure, amenities, and proximity to landmarks all had significant influences on housing price. The study indicates that a population with higher concentration of happiness correlates to higher property value and provides an innovative perspective to understand public sentiment in relation to housing price using social media data, supplemented by housing transaction data. We demonstrate a feasible framework for researchers and stakeholders to utilize in future urban and spatial geographical research.