|dc.description.abstract||The benefit of city visitors, i.e. people who live outside the city and come to visit, is that they positively affect income and employments levels in the city (van der Borg, Costa, & Gotti, 1996). Therefore, it is important for cities to develop a coherent set of goals and strategies directed towards attracting specific groups of people, rather than to promote at random (Petrisor-Mateut, Orboi, & Popa, 2013). To attain these goals cities can mobilise assets that can producing changes in the attraction and/or retention of specific segments of the population (Servillo, Atkinson, & Russo, 2011). However, to support their decisions to mobilise a certain asset, policy makers need information on how certain policy introductions or changes in the built environment, such as the placement of a new mall, will influence the attraction and/or retention of specific segments of the population.
This study shows a promising and novel approach to use mobile phone location data as a data source to empirically determine the relative importance of factors that influence city attractiveness. The method can be used for a variety of research fields including city centre retail attractiveness, urban planning and tourism destination attractiveness to determine the relative importance of factors that influence city attractiveness. The results obtained by applying our method seem logical but cannot be validated using existing literature, because our application was too broad to be able to compare it to earlier studies. Furthermore, it is very important to understand that the attractiveness of a city differs for various groups within our society (Sinkiene & Kromalcas, 2010). Therefore, this study has developed and implemented a model that can predict the most likely trip motive based on a set of trip attributes such as the arrival time and the day of week. Additionally, this study has proposed an improved method to extrapolate the sample, that is present in the mobile phone location data, to the travelling population. The main advantage of the new method is that it takes into account demographic differences of the population, which results in a more accurate representation of the actual population. The results are evaluated to the road-side measurements data by Meppelink (2016) and show a correlation ranging from .92 up to .98, depending on the situation.||
|dc.subject.keywords||city attractiveness, cities, relative importance, determinants, linear models, data mining, decision trees, statistics, classification, modelling||