Podium: Ranking Data Using Mixed-Initiative Visual Analytics
Kalidindi, Bharath V
MetadataShow full item record
Ranking points of data is utilized in everyday decision making, and multi-attribute ranking systems are a tool used to facilitate the ranking process and help make these data-driven decisions. These systems ask users to assign weights to the attributes for representing the value each attribute to a decision, which the system then uses to compute a ranking of the data. However, it is not always easy or even possible for users to quantify or understand the relative importance of each attribute to the dataset. In fact, people generally have a more holistic understanding of the data. To address these challenges, we present a visual analytic application to help people rank multi-variate data points. We developed a prototype system, Podium, that allows users to drag data points in a table to positions within the ranking they assign based on their perception of the data points value, and generate a model based on their initial ranking that represents their perception of the data. We use Ranking SVM to make these inferences and build this model that generates the attribute weights. We also present how our system can be used to understand user preferences as well as deconstruct existing rankings.