PNIELD (Projection Navigation In Extremely Large Datasets) is a framework (in development) for navigating through large high-dimensional datasets. It relies on multidimensional projections, subsampling, and user interaction. The user can zoom in to regions of interest in real-time, where the level-of-detail will be increased, as shown in the video below.
(:html:) <iframe width="560" height="315" src="https://www.youtube.com/embed/PkxZvLHPsUI?ecver=1" frameborder="0" allow="autoplay; encrypted-media" allowfullscreen></iframe> (:htmlend:)
PNIELD uses landmark projection techniques, and provides continuity to the user by keeping the positions of special landmark points fixed, in subsequent views.
Since the number of points in the projection is limited to a constant, it is visually very scalable.
See GitHub. Information on its dependencies, how to run it, and how the code is structured can be found in the README.
J. Kruiger, A. Telea, C. Hurter. Projection Navigation In Extremely Large Datasets (PNIELD). Proc. EuroVis (posters), 2017