Delta is a real-time visualization of our site structure as observed by our readership's navigation of our website.
Delta presents each reader’s page transition as a single point on a line between two NYT web pages. These web pages are embedded on a circle, with each sector representing a specific area of NYTimes.com. The duration of travel between two points is determined by the time spent on the origin page.
By showing each event as an individual point, distortions introduced by aggregation are minimized, affording a single or group of events to be placed in the context of the whole of activity on the website. Network effects regarding related content are rendered immediately visible. Emergent patterns and relationships within Delta don’t provide immediate conclusions, but a platform for us to build hypotheses and further investigate particular behaviors.
Delta also serves as a proof-of-concept for our continued investigations with streaming data. In this capacity, Delta makes use of shaders on the GPU as a means to dynamically embed and transform heavy-tailed data by altering the space in which each event is projected.