A.I. Experiments

Font Map

Using machine learning to surface new relationships between fonts.

By Kevin Ho

Font selection is one of the most common visual choices designers make—and most fall back on old favorites, or search for a font within categories. By leveraging machine learning and convolutional neural networks to spot visual patterns, Font Map helps designers understand and see relationships across more than 750 web fonts.
Built by Kevin Ho using Tensorflow and D3.js with the support of Tobias Toft, Jochen Maria Weber, and the design community at IDEO.
Submitting...
Saving...