Learnings From Yelp Network Properties

New online evaluation networks come as results of social media networks. Ebay, Amazon, Stack Overflow, and Yelp are all examples of online networks where users submit their evaluation of a particular item whether it be another user, a product, etc. These networks allow a user to submit their opinion to be read and evaluated by other users in the network. These crowd-sourced reviews act as a method for users to infer evaluations like whether a restaurant is worth going to, if a product is good quality or whether to trust an online seller. In particular, we will discuss these kinds of online networks as a network where links correspond to a rating between two nodes. We will look at the Yelp network specifically. A common model for the Yelp network is a bipartite graph. Instead we will take a novel approach of analyzing the network as triads consisting of 1 user and 2 businesses. From this proposed model, we can draw new conclusions about the underlying network structure. We analyze how status theory from the social sciences can explain the hidden principles of user evaluation. With these insights from the status theory we develop our proposed model to predict the value of links within the network. The work in this paper can be extended to several applications beyond Yelp. In general, it shows how a user’s evaluation depends on the context or the network structure it is within.