A1444
Title: Studying urban well-being through principal component analysis
Authors: Olegs Krasnopjorovs - University of Latvia (Latvia) [presenting]
Abstract: In a world displaying a plethora of information, composite social and economic indicators have become extremely popular as a way to aggregate different dimensions in a single index, dimensions that would otherwise be difficult or impossible to compare. Eurobarometer survey on the quality of life is a rich dataset containing responses for about 40 urban well-being measures collected in 83 European cities from over 70 thousand respondents. The purpose is to show how the application of principal component analysis (PCA) to the Eurobarometer survey data allows dimensionality decrease of this large dataset without a notable information loss. PCA is employed to obtain one numerical variable for each area of urban well-being (safety, trust, environment, infrastructure, public transport, governance, livability, and economic situation) and a composite numerical urban well-being indicator per city (based on the value of first principal component of the PCA), which are then rescaled to a 0-100 point scale for representation purposes. The use of PCA is justified by the Kaiser-Meyer-Olkin criterion (which has a value of 0.85 in the composite urban well-being indicator). Furthermore, the relationship between the composite urban well-being indicator and various city characteristics is explored, and its positive relationship with gross domestic product per capita in a city as well as various city amenities, as well as its negative relationship with city size and congestion are shown.
