Planets of Disparity by Liuhuaying Yang

Online social networks and information networks have become an integral part of our daily lives. However, these networks are not entirely open to everyone. Algorithms play a role in determining what we see and interact with on these platforms. They might prioritize certain people or content over others, affecting our experiences. Unfortunately, these algorithms can unintentionally exacerbate social inequalities, such as decreasing or exaggerating the visibility of minority groups and opinions.

For example, recommender systems like Who-to-Follow (WTF) tend to promote already popular users, while PageRank often perpetuates the dominance of top-ranked users or content, limiting opportunities for others. One of the reason for such biases is that algorithms heavily rely on the structural information of the network, including the directionality and number of connections between nodes (users or content pieces), or the presence of strongly connected cliques or clusters that can arise from homophily among users.

To understand how these algorithms work and their impact on different networks, we've created an intuitive tool that presents how network structure influences ranking and recommender systems. The main focus is whether a minority group is fairly represented among the top ranked users of a fictional social networks, a problem that has wide ranging implications in the real world as well. The visualization starts by explaining the essential concepts and gradually progresses to an advanced explorable dashboard, where users can freely experiment with many possible network structures and their effect on fair outcomes.

To make the abstract concepts more tangible and minimize potential disturbances from existing social biases, we use a fictional setting of "planets of disparity" in our visualization. This approach allows users to engage with the ideas more easily and gain a deeper understanding of the complex interactions within online networks.