Overlapped user-based comparative study on photo-sharing websites.

Along with the development of Web2.0 technologies and the prevalence of digital capture devices, there has been an increased popularity of various photo-sharing websites, e.g., Flickr, Instagram, and Pinterest. More and more people today are creating and sharing millions of personal photos using these photo-sharing websites. Previous studies have shown that an individual may register and participate with several online social networking websites, and is referred to as an \"overlapped user\". In this work, we use Flickr and Instagram as test platforms to conduct a comparative study on the behaviors of overlapped users on different photo-sharing websites, including their temporal distribution, location distribution, photo annotation, and social attributes. We theoretically show that observations based on the overlapped users are more significant than those that are based on independent subject groups. Moreover, the derived observations are not only helpful in understanding the multi-online social networking (OSN) participation phenomenon in the context of photo-sharing behaviors, but they provide practical instructions to photo-sharing-based user modeling applications.