This gist includes the topical descriptor ontology from MeSH (Medical Subject Headings). Here is a link that loads the data from this gist and prepopulates the cosmograph configuration
The visualization will look something like:
The layout was precomputed using the graphviz radial layout twopi after Cosmograph added custom layout support. Here is the visualization prior to the radial layout
The data is extracted from the nxontology-data mesh network. There is an intermediate private BigQuery table that was used to create the node and edge tables. The queries are below but will not work without private data access:
-- extract MeSH topical descriptors (nodes for cosmograph)
SELECT
mesh_id AS id,
mesh_label,
mesh_date_created,
mesh__depth,
ROUND(mesh__ic__sanchez__scaled, 4) AS mesh__ic__sanchez__scaled,
mesh_uri,
layout__descriptors_only__x AS x,
layout__descriptors_only__y AS y,
FROM `target-ranking.articat_latest.mesh_nxontology_node_table`
LEFT JOIN `target-ranking.tmp._dev_mesh_nxontology_node_table`
USING (mesh_id)
WHERE
mesh_class = "TopicalDescriptor"
ORDER BY mesh__depth, mesh__descendants__count, mesh_id-- extract MeSH parent-child relationships for topical descriptors (edges for cosmograph)
SELECT
relative.relative_mesh_id AS source, -- source is parent
mesh_id AS target, -- target is child
FROM `target-ranking.articat_latest.mesh_nxontology_node_table`
CROSS JOIN UNNEST(mesh__relatives) AS relative
WHERE
mesh_class = "TopicalDescriptor"
AND relative.relative_depth = -1 -- relative is parent
ORDER BY source, target