Visualizing 20,000 Books Across Six Dimensions

We tagged every book in Source Library with six independent facets — tradition, domain, form, cultural sphere, era, and epistemic mode. That created a rich dataset: 20,000 books, each with 6–12 tags from a controlled vocabulary of 68 values. Here are four ways to see the structure that emerged.

These visualizations were generated collaboratively: Derek directed the analysis and design; Claude built the code and data pipeline. The faceted classification system itself is described in Ten Thousand Years of Tagging.

What Each View Shows

Map: Semantic Similarity, Colored by Facet

Each dot is a book, positioned by semantic similarity (UMAP projection of text embeddings). Books with similar content are near each other. The toggle buttons let you recolor the same map by any facet: tradition, domain, form, sphere, era, or mode. Watch how the structure shifts — the alchemical texts (gold) cluster together when colored by tradition, but scatter across natural-philosophy and medicine when colored by domain.

Heatmap: Tradition × Domain (Bacon’s Desiderata)

The dark cells are where the collection is deep. The white cells are where it’s thin. This is Francis Bacon’s desiderata made visible — a map of what knowledge is missing. Alchemy × natural-philosophy is dark (hundreds of books). Sufi × music is white (a known gap). Every white cell is a curation opportunity.

Flow: How Traditions Feed into Domains

A Sankey diagram showing how intellectual traditions flow into subject domains. Some traditions are narrow — masonic texts flow almost entirely into theology. Others are broad — classical Greek and Roman thought feeds into history, ethics, natural philosophy, theology, and art. The width of each ribbon is the number of books at that intersection.

Connections: When Traditions Overlap

A chord diagram showing tradition co-occurrence. When two traditions appear as tags on the same book, they share a ribbon. Hermetic and alchemical are tightly linked (many books carry both tags). Classical and neoplatonic overlap heavily. But vedic and rosicrucian never appear together — they are intellectually isolated from each other in this collection.

What Surprised Us

The collection is 60% treatises. The form distribution is dominated by sustained arguments on a single subject. Anthologies, manuals, and poetry are a distant second. Source Library is overwhelmingly a collection of thinkers thinking.

Christian mystical is the second-largest tradition after classical Greek & Roman. We thought of the library as primarily “esoteric” — alchemy, Hermeticism, Kabbalah. But the data says it’s as much a library of Christian contemplative theology as it is of Western esotericism.

The Latin West dominates. Despite hundreds of Chinese, Sanskrit, and Arabic texts, the cultural sphere distribution is heavily Latin + Vernacular European. The non-Western traditions are present but thin. The heatmap makes this painfully clear.

Speculative mode is everywhere. Over half the books are tagged “speculative” — reasoning from principles rather than from observation. This is a library of metaphysics, not of experiment. Bacon would have been disappointed.