Identifying and Predicting Neighbourhood Level Gentrification: A Data Primitive Approach

Identifying and analysing neighbourhood change is a critical task for urban planners and policy makers and is an active academic field. However, traditional approaches to neighbourhood change often rely on temporally static data and methods that reduce complex processes to one cluster label, or one score for example. This leads to a fragmented understanding of neighbourhood dynamics, on a temporal scale that does not align with the processes, resulting in the failure to capture their complex and multifaceted nature. These limitations highlight the importance of adopting new and innovative methods to provide more accurate and dynamic insights into neighbourhood dynamics. This research subsequently proposes a new approach, data primitives, and a methodological framework for their application. Data primitives are measurements of the fundamental components that capture the driving characteristics of clearly conceptualised neighbourhood processes. Their utility is explored in a regional analysis, identifying 123 cycles of gentrification and their respective temporal properties, which are exhaustively validated via Google Earth and Google Street View. This demonstrates the effectiveness of data primitives at capturing processes, and quantifying their changes over time, to provide a more comprehensive picture of neighbourhood change. These validated cycles of gentrification are used as a training dataset for training three machine learning algorithms for predicting gentrification in England. Three models were created to predict the presence of gentrification, the type of gentrification, and the temporal properties of the predicted types of gentrification in England. These predicted cycles of gentrification are explored, generating novel insights for the neighbourhood change and gentrification communities. Overall, the results of this research have important implications for urban planning and policy making, as they can provide a framework for informing decisions on where to invest resources and how to mitigate the potential negative effects of gentrification, in an appropriately scheduled timetable of interventions. They also provide a framework for uncovering novel insights into the complexities of neighbourhood processes, and their impacts upon neighbourhood change, thus developing upon knowledge in suitable academic fields.