Less is more: empirical design criteria for a tourist place recommendation service which decelerates the visiting experience

Abstract

Tourist recommendation services provide suggestions for places of interest. While the spatial structure of a tourist's visit has been successfully exploited to improve the quality of place recommendations, its temporal structure has not been taken into account so far. This is surprising because the importance of the visitors' time management is well known from tourism research. We present findings of an empirical study that sheds light on the temporal aspects of tourist exploration behaviour and discuss implications for the design of tourist-recommender systems. The data-set consists of interviews, GPS tracks and geo-referenced photo sequences from visitors of a typical Middle European destination for cultural tourism. Among the results relevant to place recommending are the following: (1) the set of places is often replanned during the visit; (2) comparing place popularity based on photo frequency and place popularity based on staying time, we found notable differences in rank ordering; and (3) the photo frequency and the detour sinuosity of the GPS tracks are decreasing slowly over time. We interpret this last finding as a time-geographic cone effect. Based on the empirical results, we discuss the implications for the design of tourist-recommending services and propose a corresponding user interface utilising a recommendation strategy that tries to counterbalance the time-geographic cone effect, that is, to decelerate the visiting experience.