§ Method

How the record was read

This site reads a single database built by the insight-bridge pipeline from its corpus: 999 submissions, 1,293 submission documents, 11,649 passages, 1,737 submission–topic memberships, 1,720 submission-level stances, 312 sector perspectives.

Everything on every page is a live query against Cloudflare D1 (the run's own tables, copied whole) and Vectorize (the embeddings), through fully-typed Drizzle. The same data is exposed to agents over an MCP server at /mcp.

What was done, in order

Each step below is a stage this run actually executed, read from the run's own record of itself; a stage the run was configured to skip is not listed.

  1. Collection

    The published submissions to the 2026 Murray–Darling Basin Plan Review (consultation 5 Feb – 1 May 2026) are collected, split into documents, chunked and normalised. Submitting entities carry their published facets: sector, respondent type, named status, ICIP flag.

  2. Reading in the submission documents

    Every submission document was stored whole under its submission, its text cut into passages small enough to compare — 11,649 of them across 1,293 submission documents from 999 submissions — and each passage embedded, which is what makes a search by meaning, and the grouping below, possible at all. 426 no-content stubs, submission records with no retrievable text, are not in the record, so document counts describe what was actually read.

  3. Reading each submission document end to end

    A language model read every submission document in full against this run's profile and recorded what it found as key points, each with the verbatim quotes that support it: 9,238 key points and 24,179 quotes. Every quotation is verbatim from the source submission document: each is verified against the submission document's text before it is stored, and one that cannot be found there is never presented as verified. 98.1% verify, and a quote that does not is never shown here. Each document was also tagged across four controlled vocabularies (Basin policy theme, Evidence type, Stakeholder interest or role, Geographic focus) and given a recommendations-and-tradeoffs analysis under four fixed subtopics — concrete actions requested, conditions attached, responsible actors, outcomes sought or feared.

  4. Grouping the passages into topics

    Topics are found from the passages themselves, bottom-up, by grouping the passages across the corpus that say the same thing; nothing supplies a list of themes beforehand. A topic has to be raised by more than one submission to form: one submission's volume cannot make a topic on its own, and the rare topic that only one submission carries is flagged as such. The run did not record the settings it clustered with. Each passage, and each submission, is then placed against every topic at a graded strength (exemplar, high-value, member), so a submission that touches a topic in passing is recorded as well as the submissions that define it, and a passage can belong to more than one topic. 86 topics resulted, over 8,516 passages.

  5. Writing each topic up and reading positions against it

    Each topic is written up as a proposition with key points, and every exemplar and high-value submission's position is recorded against that proposition, on this run's own scale, Supports, Builds on, Mixed, Redirects, Opposes and Unclear. Positions are relative to the topic's own framing, never absolute agreement, and every submission is measured against the same proposition. 1,720 positions were written, and the same was done for each value of source sector: 312 perspectives. A position here records how a submission stands against one topic's particular framing, and is never a verdict on the subject in general.

  6. Offering the topics upward into themes

    The topics were grouped again, generation by generation, each generation clustering the one below it; the settings it ran with were not recorded. This site reads the tree as the run left it — here, 20 families above the topics, each named by the model from its members.

  7. Placing the topics in a sky

    Each topic's place in the galaxy was computed once and stored with the run, seeded so it is the same sky every time. That is what lets these pages draw the constellation as plain graphics, with no WebGPU and nothing solved in your browser.

  8. Loading this site

    The run was made in July 2026 by an earlier version of the pipeline, and was carried onto the current one in October 2026 as it was written: no step was re-run and no text changed, and its quotes were re-verified by the pipeline’s own rule. Outside the pipeline, before loading: the run database was copied whole to the cloud — every table as the pipeline wrote it, with the passage embeddings moved to a vector index — and two facts were derived from it for these pages, each document’s and each passage’s dominant topic. Nothing was flattened or re-classified here; every figure on every page is a live query over the run’s own tables.

What each submission was classified by

Every facet the run's profile declares. A facet marked compared is one the analysis wrote a perspective for, value by value — the lens you can read a topic through. Facets that are not compared ride beside it as facets of the record.

Sector
Compared. Irrigated agriculture · Environment and conservation · Other · Government · Community and regional development · Dryland agriculture · Research and academia · Traditional Owner, Aboriginal or First Nations · Recreation · Unknown Left out of comparisons: Unknown.
Respondent type
Recorded as written, with no fixed list of values. Left out of comparisons: Unknown.
Submission modality
Recorded as written, with no fixed list of values. Left out of comparisons: metadata_only.
Source-document type
Recorded per submission document. Recorded as written, with no fixed list of values.

What a position means here

Read weakest claim first. Supports, Builds on agree with a topic's proposition; Mixed, Redirects, Opposes push back against it; Unclear reads as neither.

Supports
accepts the cluster's principal proposition; minor caveats, concerns or suggested details remain Supports
Builds on
accepts the proposition's core and makes a concrete, material contribution that extends, strengthens or develops it; do not use Builds on merely because the source mentions a caveat or change
Mixed
gives substantive reasons both for and against the proposition without a clear overall direction
Redirects
does not accept the proposition as framed but constructively argues for a materially different approach, priority or framing
Opposes
clearly rejects the proposition or argues against pursuing it
Unclear
there is not enough evidence to determine the position

Position is relative to the cluster proposition, not the source's tone. Nuance alone is not Mixed, and constructive engagement alone is not Builds on.

How to read the output

These rules apply to every Insight Bridge corpus. They are properties of the method, not caveats about a particular run.

  • Positions are relative to a proposition. A source’s position records how it stands against that cluster’s particular framing — not whether it agrees with some absolute claim. The same source can support one cluster and redirect a neighbouring one that covers similar ground differently.
  • Propositions are written from the cluster’s own members. Because the argument is built from the sources that were grouped together, and positions are read relative to it, a degree of agreement is built into the method. Comparisons between groups carry weight; a corpus-wide agreement rate does not.
  • Counts describe the corpus, not the world. Every corpus here is curated. “N sources say X” measures what was collected and is never a measure of how common X is in the field.
  • Clusters differ in how many distinct sources back them. A long document can fragment across many clusters, so weight a theme by the distinct sources beneath it rather than by how many clusters it contains.
  • Every extraction and position is a model judgement. Key points, stances, propositions and syntheses are produced by a language model reading the source. They inherit its calibration and are not determinations of fact.

In this run

Not analysed. ; ; ; ; ; ; ; ; ; ; ; ; ; ; ; ; ; ; ; ; ; ; ; ; ; ; ; ; ; ; ; ; ; ; ; ; ; ; ; ; ; ; ; ; ; ; ; ; ; ; ; ; ; ; ; ; ; ; ; ; ; ; ; ; ; ; ; ; ; ; ; ; ; ; . Step 1 produced no key points for these documents; they keep their topic memberships.

Kept out of the analysis. submissions whose icip_content is Yes. They are in the record and can be read, but no topic or perspective was written from them.

426 no-content stubs, submission records with no retrievable text, were excluded at load, so document counts describe what was actually read. ICIP-flagged submissions (18, those declaring Indigenous Cultural and Intellectual Property) keep their entity and document metadata so the record stays honest, but their verbatim text — chunks, quotes, key points — is withheld from this deployment pending ICIP consent processes.

Where the disagreement is. The framings here are things the submissions actually argue — "halt water buybacks", "recover water for the environment" — and 89% of entity stances read supportive against them, in a corpus containing a deep two-bloc split. Dissent therefore surfaces between opposing framings rather than inside one topic's rows; the Fault lines view is built to show it.

What else was asked of every submission document

Beyond the standard reading, this run declared its own questions and put each of them to every submission document.

Basin policy themes
Tag every policy theme substantively discussed in this source document. Use High relevance for a central argument or recommendation, Medium for material supporting discussion, and Low only when the theme is mentioned with some analytical significance. Do not tag themes merely appearing in form headings or quoted background.
Basin context tags
Classify context expressed in the document. Apply multiple tags where warranted and use only evidence in the document. Do not infer a stakeholder identity solely from a policy position.
Basin recommendations and tradeoffs
Extract the document's concrete recommendations and policy positions without judging whether they are correct. Preserve important conditions, trade-offs, responsible institutions, intended outcomes and exact supporting quotations.

This run kept no record of the models or settings it ran with, so this page names none of them. Everything else here is read from the run itself.

999 submissions1,293 submission documents426 stubs excludedICIP text withheldD1 + Vectorize11,649 passages9,238 key points24,179 quotes · 98.1% verified86 topics · 0 removed20 families1,737 memberships1,720 stances312 sector perspectives