Mining business competitiveness from user visitation data

Abstract

Ranking businesses by competitiveness is useful in many applications including business (e.g., restaurant) recommendation, and estimation of intrinsic value of businesses for mergers and acquisitions. Our literature reveals that previous methods of business ranking have ignored the competing relationship among businesses within their geographical areas. To account for competition, we propose the use of PageRank model and its variant to derive the Competitive Rank of businesses. We use the check-ins of users from Foursquare, a locationbased social network, to model the winners of competitions among stores. The results of our experiments show that Competitive Rank works well when evaluated against ground truth business ranking.