Investigating the Utility of the Weather Context for Point of Interest Recommendations

Abstract

Point of interest (POI) recommender systems for location-based social networks, such as Foursquare or Yelp, have gained tremendous popularity in the past few years. Much work has been dedicated to improving recommendation services in such systems by integrating different features (e.g., time or geographic location) that are assumed to have an impact on people’s choices for POIs. Yet, little effort has been made to incorporate or even understand the impact of weather on user decisions regarding certain POIs. In this paper, we contribute to this area of research by presenting the novel results of a study that aims to recommend POIs based on weather data. To this end, we have expanded the state-of-the-art Rank-GeoFM POI recommender algorithm to include additional weather-related features such as temperature, cloud cover, humidity and precipitation intensity. We show that using weather data not only significantly improves the recommendation accuracy in comparison to the original method, but also outperforms its time-based variant. Furthermore, we investigate the magnitude of the impact of each feature on the recommendation quality. Our research clearly shows the need to study weather context in more detail in light of POI recommendation systems. This study is relevant for researchers working on recommender systems in general, but in particular for researchers and system engineers working on POI recommender systems in the tourism domain.