Analysis of Bidding Networks in eBay: Aggregate Preference Identification through Community Detection

Abstract

Statistical analysis of networks plays a critical role in the context of economics and the social sciences. Here we construct a bidding network to represent the behavior of users of the eBay marketplace. We study the eBay markets for digital cameras and liquid crystal display screens, and employ network analysis to identify aggregate structure in bidder preferences. The network that we construct associates auctions with nodes, and weighted edges between nodes capture the number of bidders competing in a pair of auctions, where said bidders ultimately win in only a single auction. We show that current community detection methods applied to this network allow for the identification of goods that are considered substitutes and complements, and thus the identification of aggregate preference information. In closing we suggest additional opportunities as well as challenges for the analysis of structured data in electronic markets.