Discovering the impact of urban regeneration on place identity change with user-generated contents analytics

Abstract

What makes a city a unique place and experience? This question has long been asked in the field of urban design in relation to successful placemaking with strong identity. In this regard, measuring how people understand the urban environment has been an important theme in urban research. Over the recent decades, the shift to regeneration strategies in urban policies has been demanding the need to discuss regeneration in relation to the place identity concept considering changes in the urban environment. In the era of urban information systems and big data, user-generated contents (UGC) are opening new visions for identifying the meaningful semantics of people-place relationships. Within this context, two research problems are defined: (1) place identity extraction at a collective scale

and (2) verifying place identity change in urban change scenarios.The overall methodology of this dissertation is organized based on the phases of the UGC analytical framework - retrieval, processing, analysis, and implication.

To address the research problems, we focus on the research potential of non-georeferenced texts from social media and attempt to extract geographical context and place-related semantics for place identity assessment using three regeneration projects in Seoul, Korea as study sites. This dissertation proposes a novel method to construct a crowdsourced place identity dataset from which people’s collective understanding of places are represented. Implications on place-specific meanings of each site are found through initial data analysis, which were further investigated in later analyses. First, we apply biterm topic models on non-georeferenced texts to discover subjective information that can explain the identity of regeneration sites. We analyze the discovered topics with a longitudinal approach and compare their spatiotemporal hotspots with the timeline of regeneration schemes to reveal the congruence between place identity development and urban change scenarios. In addition, they are found to have positive correlation with commercial and tourism revitalization indices at each site. Second, we explore the driving factors of place identity intensity (PII) by adopting a local regression approach. Specifically, we apply geographically weighted regression models to uncover the spatially variability of regeneration factors in place identity.

The results confirm that place identity is rather dynamic than static, and should be considered within the changing urban context. In particular, this has been found to change in close connection with regeneration plans. Moreover, local regression results demonstrate counterfactual findings in regard to urban regeneration and place identity. For instance, we contradict the concerns of redevelopment efforts for being disruptive to place identity but instead show that physical change may reinforce place identity in commercial and residential areas. In addition, we suggest counterintuitive place identity-oriented strategies to attract new businesses in regeneration sites. In consequence, the findings provide evidence of commonly raised criticisms of urban regeneration approach - commodification of place, gentrification, and the myth of its comparative advantage over redevelopment strategy - and calls for an identity-driven approach in future planning.