Applying non-parametric models to explore urban life satisfaction in European cities

Abstract

Although urban quality of life is one of the important topics in most of the European policies, these policies are not supported by empirical analysis. Furthermore, there are few studies in urban quality of life area that consider a sufficiently large number of cities and related factors and those that focused on European cities applied parametric analysis techniques that cannot handle multicollinearity. To remedy the aforementioned shortcomings, in this study, a large number of factors related to urban quality of life are investigated in different European cities, and to handle multicollinearity, non-parametric analysis techniques are adopted. The data stem from the Eurostat (2015), collecting information on urban quality of life for > 40,000 citizens in 112 urban areas. Different non-parametric modeling techniques are applied and the results of the method that yielded the highest overall accuracy, i.e. the C5.0 algorithm, are presented. The results show that from the wide scope of considered factors, five main variables play an important role in urban life satisfaction, i.e. (i) feeling safe in the city, (ii) satisfaction with healthcare services in the city, (iii) satisfaction with the state of streets and buildings in the neighborhood, (iv) satisfaction with public transport in the city, and (v) availability of retail shops.