Spot: Locating social media users based on social network context

Zhi Liu, Longbo Kong, Yan Huang

Abstract: A tremendous amount of information is being shared every day on social media sites such as Facebook, Twitter or Google+. But only a small portion of users provide their location information, which can be helpful in targeted advertisement and many other services. In this demo we present our large-scale user location estimation system, SPOT, which showcases different location estimation models on real-world datasets. The demo illustrates three algorithms—friend-based, social-closeness-based, and an energy & local social coefficient model—along with a confidence-based iteration method, visualized via Google Maps with dynamic statistics (accuracy, average friends used). Two datasets (Twitter: 148,860 users; Gowalla: 99,563 users) demonstrate that close friends yield better estimates.