The org as a network
Every human contributor connected to every repository they have committed to, in one picture. Bus-factor numbers say concentration exists; this page shows where the thin threads are — a repository whose only line reaches one person is a single point of failure drawn literally. Source data: contributors.csv (alias-merged upstream), bus-factor classes from bus_factor.csv (reports/bus_factor.md).
Repositories
Human contributors
Repos reachable from one person only
Edges (person→repo)
Force map
How to read: Squares are repositories, colored by bus factor (red = one person carries the majority of the work, amber = two, green = three or more); circles are people, sized by their share of all contributions. A line means that person has commits in that repository; line thickness scales with how many. Example 1: A red square dangling from a single line is the sharpest continuity risk in the org — one account lost, and that repository's working knowledge goes with it. Example 2: If any square were green (bus factor ≥ 3), it would mark a repository with genuinely shared ownership — as of this snapshot, not a single one of the 76 repositories qualifies.
Conclusion: The picture is a hub-and-spoke system, not a mesh: three large circles hold nearly every repository, 65 of 76 squares are red (single-majority ownership), the remaining 11 amber, and not one repository reaches bus factor 3. There is no path between most repositories except through funderburkjim, drdhaval2785, or gasyoun; recruiting even one contributor into the rim would visibly re-wire this graph — precisely the bus-factor mitigation the Community page quantifies.
Adjacency matrix (accessible view)
How to read: The same data as the force map, as a matrix: one row per contributor (top 10 by volume), one column per repository, a cell where that person has commits, darker for more. This view is exact where the force layout is impressionistic — use it to look up who has touched what. Example 1: A row that spans most columns is a generalist maintainer; a row with one dark cell is a specialist. Example 2: Columns with exactly one filled cell are the single-point-of-failure repositories from the map, now individually identifiable.
Conclusion: Only the top three rows have meaningful width; rows four to ten are narrow specialists. Reading down almost any column lands on the same one or two names — the matrix confirms that the network's apparent breadth is three people's reach, not a distributed community.
The cross-repo issue-reference graph (which issues cite which across repositories) needs issue bodies the committed snapshot does not carry; it is tracked as an API-gated extension under Workstream G4 in the roadmap.