Family Tree
Genealogy and relationship-graph experiments — trees, DAGs, and hierarchy layouts in the same dark predictive site language.
A family tree is a claims hierarchy with better stories: nodes, edges, and the discipline of not inventing parents you cannot evidence.
What’s here
On this page
The notes below are the published version. There is no separate project folder on this site.
Graph thinking
Pedigrees, org charts, and claim/policy hierarchies share the same grammar: directed edges, cycles you must forbid, and missing-link honesty.
Viz craft
Force layouts look clever; layered trees read. Prefer readability when the audience is non-technical family (or auditors).
Practical notes
- Identity first. Deduplicate people before drawing edges. Fuzzy name match without dates is how you invent cousins.
- Evidence tags. Attach source notes to edges (census, certificate, oral history). Graphs without provenance age badly.
- Cycles mean bugs. In a pedigree DAG, a cycle is a data error — treat it like a failed unit test.
- Reuse elsewhere. The same layout code informs network viz for movies, treatment pathways, and org risk maps.
These notes are the published family-tree page. The graph rules below are the part worth keeping.