Trajectory Fluid

Donghao Ren

John O'Donovan

Network data is ubiquitous. This installation presents an artistic approach towards understanding how people perceive and analyze network data in different visual representations. Mouse trajectories from a large dataset collected from 600 Amazon Mechanical Turk (a crowdsourcing service) participants is presented. The trajectories are revealed in both the screen space and the abstract space. In the screen space, trajectories are presented with real-time visualization. In the abstract space each trajectory is mapped to a single 2D point thus aiming to capture the similarities and dissimilarities between trajectories. The abstract space is computed using the t-SNE (t-Distributed Stochastic Neighbor Embedding) algorithm with the DTW (Dynamic Time Warp) distance. Three-dimensional Hermite splines and semi-transparent bands are drawn to connect the screen space and the abstract space. Mouse movements and annotations in the screen space generate turbulence that propagates to the abstract space among the lines, which resembles a “fluid” from the screen space to the abstract space.

Project Image