Automated Location-Aware Influencer Evaluation

Abstract

Location-aware influencer is a user who caters to a geographically specific audience. Examples of location-aware influencers are the local news reporter, the city mayor, local business and others. Such influencers can be identified by the location that captures the biggest percent of the influencer's followers.Size of online social networks (OSNs), their API data limits, and lack of precise location data make it a challenging and timeintensive process to identify such influencers. This paper proposes an algorithm to establish, validate, and continuously update a repository of location-aware influencers via targeted collection.The evaluation of location-aware influencers is typically limited to small handlabeled datasets, but automating key steps allowed us to evaluate thousands of influencers. Our approach is verified using multiple datasets including influencers associated with city level locations via automatic Google searches. A repository of such location-aware influencers is of interest for content recommendation and for studying location related communities from influencer's followers. The approach is developed over Twitter but applies to other social networks.