Affiliations & Networks

Person-to-organisation associational memberships: schools, universities, churches, fraternal orders, private clubs, think tanks, industry boards, and advocacy groups for politicians and public figures tracked in this platform.

Status: staff-only until data verification is complete. The public launch is gated on manual review of every row.

What it shows

• Total affiliations — count of verified person → organisation links.
• Subject coverage — what fraction of tracked subjects have at least one recorded affiliation.
• By organisation type — breakdown across schools, universities, religious institutions, fraternal orders, clubs, think tanks, industry bodies, boards, and advocacy groups.
• By confidence tier — confirmed (primary source), reported (credible media), claimed (self-declared), alleged (single source, flagged).
• Top organisations — organisations ranked by member count, linking to per-organisation detail pages.
• Shared affiliations — pairs of subjects linked to the same organisation. This drives medium-severity "network proximity" flags on the Relationships dashboard. Shared membership is not proof of wrongdoing — only worth examining.

How to read it

Use the Top Organisations panel as an index into the alumni/club/board networks. Click through to an organisation detail page to see every affiliated subject, their role, period, confidence tier, and source link.

Use the Shared Affiliations panel to spot where two or more subjects converge on the same organisation — this is the data that makes interlocks visible (e.g. the same industry board linking a minister and a tenderer's director).

Data sources

Every affiliation row has a required DataSource foreign key. The enrichment pipeline's first-pass parser reads the Wikipedia article already fetched by enrich_from_wikipedia and extracts education, religion, and club/board mentions using conservative regex patterns. Admin entries supplement this with primary sources.

Planned additions: targeted Hansard, ABC, Guardian, AFR, Age, SMH, Saturday Paper and Crikey searches for "<name>" "alumni", "<name>" "parishioner", "<name>" "board of", "<name>" "member of" patterns, reusing the existing enrich_from_news machinery.

Source policy

The following source categories are excluded at ingest time via AFFILIATION_SOURCE_DENYLIST in apps/projects/tasks.py:

• Milton William Cooper material and sites that syndicate it (e.g. jewworldorder.org).
• Anonymous "truth/unveiling/exposed" blogs with no named author or editorial process.
• Video channels (Odysee, Telegram, etc.) whose editorial framing rests on ethnic, religious, or conspiracy narratives.

Rationale: these sources are unfalsifiable, create defamation exposure, and would destroy the credibility of the genuine accountability work in this platform. A denylist entry can be added or removed only through discussion in code review — do not edit the list without raising it.

This is stricter than the CLAUDE.md-level "every fact must be attributable" rule, because the subject matter (religious affiliation, club memberships, alleged fraternal ties) is historically the most contested and most weaponised.

Confidence tiers

Caveats

• Shared membership is network proximity, not proof of wrongdoing.
• Religious affiliation is often transient and private — treat reported tier claims with care.
• Alumni lists can be decades old — is_current is usually False for education-type affiliations.
• The Wikipedia-based auto-extractor is conservative but not perfect. Review any automatically-created Organisation row before it goes public.

Related pages

• Relationships — the existing person-to-person relationships dashboard.
• Network Graph — Cytoscape visualisation; staff users can enable Organisation nodes with ?affiliations=1.
• Mates Detector — adjacent network-proximity analysis.

Last reviewed: 11 Apr 2026