Characterizing popularity dynamics of online videos
Foursquare, the currently most popular location-based social network, allows users not only to share the places (venues) they visit but also post micro-reviews (tips) about their previous experiences at specific venues as well as "like" previously posted tips. The number of "likes" a tip receives ultimately reflects its popularity among users, providing valuable feedback to venue owners and other users. In this paper, we provide an extensive analysis of the popularity dynamics of Foursquare tips using a large dataset containing over 10 million tips and 9 million likes posted by over 13,5 million users. Our results show that, unlike other types of online content such as news and photos, Foursquare tips experience very slow popularity evolution, attracting user likes through longer periods of time. Moreover, we find that the social network of the user who posted the tip plays an important role on the tip popularity throughout its lifetime, but particularly at earlier periods after posting time. We also find that most tips experience their daily popularity peaks within the first month in the system, although most of their likes are received after the peak. Moreover, compared to other types of online content (e.g., videos), we observe a weaker presence of the rich-get-richer effect in our data, demonstrating a lower correlation between the early and late popularities. Finally, we evaluate the stability of the tip popularity ranking over time, assessing to which extent the current popularity ranking of a set of tips can be used to predict their popularity ranking at a future time. To that end, we compare a prediction approach based solely on the current popularity ranking against a method that exploits a linear regression model using a multidimensional set of predictors as input. Our results show that use of the richer set of features can indeed improve the prediction accuracy, provided that enough data is available to train the regression model.