We present the concept of a text-based topic similarity measure to be used towards content-based recommendations in visualization repositories. We investigate four applicable NLP models and conduct a user study that demonstrates it is possible to obtain good alignment between human similarity perception and off-the-shelf model predictions.

The VizSnippets computational pipeline compresses the visual and textual content of bundles into representative previews that is adaptive to a provided pixel budget and provides high information density with multiple images and carefully chosen keywords.