DD Visual Intelligence Research Protocol (VIRP)
Version 1.0 — effective 19 November 2026. Applies to all DD Visual Intelligence editions, editions in development, and all future editions in the series.
Source hierarchy and categories
Sources are classified into five categories. The category determines how a source may be used, not how it is ranked.
Primary
Official documents: government websites, ministry records, legislation, strategic frameworks, procurement records, official statistics. S1 sources are directly citable as the sole basis for a claim.
Institutional
International organisations: UN agencies, OECD, EU institutions, multilateral bodies. Citable in conjunction with at least one S1 source for the same claim, or as the sole source for a claim about that institution itself.
Corroborating
Academic publications, think tanks, specialised journalism, expert testimony. Never the sole source for a claim. Used to provide context, corroborate or qualify claims already supported by S1 or S2.
Secondary
Media coverage, aggregators, databases and indices. Discovery only: may be used to locate primary sources, never cited as a claim source.
Speculative
Social media, AI chat interfaces, forums and unverified commentary. Excluded from research. Social media content by the government or institution in question may be treated as S1 when published from an official account.
Every source used in an edition is recorded in the source appendix with category, institutional author, title, publication or modification date, URL, access date and relevance note.
Research phases
Collection
Open-source search across all applicable source categories. Sources are recorded with category, institutional author, title, publication or modification date, URL, access date and relevance note.
Verification
Each claim is traced to the closest authoritative source. Claims that cannot be verified at S1 or S2 level are excluded from the visual. The absence of public evidence is never treated as proof that a capability or practice does not exist.
Assembly
Visuals are constructed from verified claims only, with their evidence status made explicit. A visual is only published when it meets the visual-argument standard.
Evidence categories
Every claim in a Visual Intelligence edition carries one of three categories. The category determines the claim's visual weight and the language in which it may be stated.
Directly evidenced by an S1 source, or by an S2 source in conjunction with an S1 source. May be stated as a direct assertion.
Inferred from multiple sources under DD editorial judgement, where no single source states the claim directly. Labelled as DD Classification. Never stated as fact.
DD interpretive analysis based on documented evidence. Labelled as DD Analysis. Must reference the evidence on which it is based.
DD Classification and DD Analysis carry visible labels on the edition page and in the source appendix.
Visual-argument standard
A visual is publishable when all of the following are true:
- It makes one argument.
- Every data point is traceable to a recorded source.
- A reader can reconstruct the reasoning from the visual and the source appendix alone.
- The visual does not imply evidence beyond what is recorded.
- The accessible text equivalent conveys the same argument as the visual.
A visual that cannot meet this standard is not published. A research question that cannot be answered with available evidence is recorded as unanswered, not answered speculatively.
Corrections and versioning
Editions are versioned. Corrections, additions and significant new evidence produce a new version with a dated entry in the public version history. The canonical URL remains stable across versions.
When a correction materially changes the finding of a published edition, a correction note is displayed at the top of the edition page.
An edition that is superseded by a more complete treatment of the same research question is marked Archived. Archived editions remain accessible and indexable at their canonical URL, clearly labelled, and are not deleted.
Visual and accessibility standards
All Visual Intelligence editions follow these standards without exception:
- Colour is never the only carrier of meaning: every colour-coded element has a text label, a pattern or a position that conveys the same information.
- All visuals have an accessible text equivalent that conveys the visual argument in full.
- All text in visuals meets WCAG AA contrast requirements against its background.
- Visuals are legible at 320 px viewport width without horizontal scrolling.
- Charts and maps are presented in vector format (SVG) where possible. Raster formats are used only where vector is not achievable.
- Evidence-status labels (Fact / DD Classification / DD Analysis) are visible on the visual itself, not only in surrounding text.
Scope and limitations
Visual Intelligence examines publicly available evidence. It does not assess classified activity, internal processes not publicly documented, or capability that a government has not chosen to make public.
The absence of public evidence for a claim is recorded as an absence of public evidence. It is not treated as proof of absence.
Editions state their scope explicitly — the countries, institutions, domains or time period examined — and do not generalise beyond it.
Visual Intelligence does not rank countries, ministries or institutions. Where comparisons are made, they compare institutional patterns and evidence, not performance.
Relationship to other DD research
Visual Intelligence operates under the same editorial standards as all DD research. Its evidence discipline is aligned with the DD Country Profile Methodology. Its design standards are governed by the DD Design System.
Visual Intelligence is a distinct research format. It is not a summary of Country Profiles, a visualisation of Insights articles, or a derivative of DD Frameworks. It addresses research questions that are best answered visually.
- Document
- DD Visual Intelligence Research Protocol (VIRP)
- Version
- 1.0
- Effective
- 19 November 2026
- Applies to
- All DD Visual Intelligence editions
- Maintained by
- Diplomats.Digital