.jpg)
Gentrification perpetuates impacts across society and space which are both tangible and non-trivial in nature. Thriving on the widening socio-spatial inequalities in our present day and age, gentrification's pervading presence have grown ever stronger and in increasingly varied forms. Yet, gentrification has commonly been conflated with other forms of neighbourhood ascent, while gentrification's nuanced trajectories have in themselves been indifferently treated. Such trends have consequently limited urban planners and city administrators in their capacities to formulate effective measures to tackle gentrification's festering problems. Our research was hence build upon the ambition to critically decipher the dynamics of neighbourhood change and distinguish between the differing forms of gentrification at the Lower Super Output Area (LSOA) scale across London. By visualising our study's results in the digital maps below, it is hoped that new insights and knowledge can be drawn, which will eventually contribute towards robust policy-making and solutioning efforts aimed at quelling gentrification's harmful effects.
Relative ascent, decline and stability constitute the broadest level of the neighbourhood change schema. Applying PCA on a set of proxy variables (i.e. house prices, income, % of residents classified as NS-SEC 1 and with higher qualifications), a Composite Index that broadly quantified the states of LSOAs was derived. Neighbourhood change was then established based on relative differences in a LSOA's Composite Index scores / rankings between 2001 and 2011.
Dis-aggregating the divergent trajectories of neighbourhood ascent and gentrification is critical in gaining clairity on the urban processes active in localised neigbourhoods. Targeted iterations of K-Means clustering was done to obtain the dominant neighbourhood change and gentrification typologies active across London's urban landscape, and their corresponding typology characteristics.
Given gentrification's inherently spatial nature, geospatial methods are ideal in eliciting deeper geographic insights into the urban phenomemon. Employing Hotspot Analysis on the varying manifestations of gentrification, the hotspots of LSOAs undergoing super-gentrification, marginal gentrification and mainstream gentrification, and their spatial underpinnings, were identified and visualised.
Machine Learning is a specialised field of Artifical Intelligence that is quickly gaining popularity as a robust empirical method for undertaking research. Specifically, a Random Forest multi-classifiication model was trained and calibrated using the outputs of the K-Means clustering and geospatial analysis to determine whether a LSOA was gentrifying, and if so, their associated gentrification typologies. By feeding the model with recent data from the latest census and other relevant datasets, predictions of gentrification's future frontiers were generated.
Navigating the website
- Access the section of interest by selecting the icons in the navigation bar towards the right of the screen
- Additional descriptions for the Carto visulaisations can be viewed by clicking on the blue-coloured banners along the left of each map
- Interactive charts and graphs are displayed on the widgets integrated with each visualisation
References & Credits
Quote on homescreen: Lees, L., Slater, T., Wyly, E.K., 2008. Gentrification. Routledge, New York, NY.
Icons in the navigation bar made by smashicons, freepik, geotatah & dave-gandy from www.flaticon.com