← Port Analytics

Port Network

Port-to-port traffic as a network graph -- which ports are structural chokepoints vs. just busy, from real vessel transitions between consecutive port stays.

Nodes are DfT major-port groupings, edges are "a vessel's next port stay after leaving A was at B" (weighted by distinct vessel count, self-loops dropped), built from weeks 2026-W37, 2026-W38 of the permanent ais.port_weekly_visit archive. No max-gap cutoff between consecutive stays -- a vessel that went dark for months between two visits still counts as a direct transition; revisit if this turns out to matter once more weeks accumulate. Regenerate via tools/build_port_network.py. Last built 2026-09-21T01:05:04.374692+00:00.

- -
Click a port to see its strongest connections
No port selected.

Busy, but not a real connector

High raw-visit rank, low betweenness rank -- traffic terminates here rather than passing through.

Hidden structural connector

Low raw-visit rank, high betweenness rank -- more structurally important than traffic volume alone suggests.

Port Raw visits Raw rank Betweenness Betw. rank PageRank