A regulatory-grade monitoring instrument, built from structured facts — not sentiment. This is the visual grammar by which we illustrate a stack-ranking system as it observes events, flows of value, and competitive posture across a vertical.
Every report ShurIQ produces is an observation against a rubric. The rubric defines the ideal state for each category we rate; the ontology defines the universe of entities, events, and flows we are allowed to observe; the knowledge graph makes that universe queryable. This document is the first canonical walk-through of how those three pieces compose — using the vertical drama sector (Chinese-origin micro-series apps) as the worked example.
What follows is not a deck. It is the operating grammar — the set of primitives, transforms, and conventions we use to depict what ShurIQ is doing at any given moment. When a client sees a ShurIQ report in production, every mark on the page traces back to one of the primitives defined here.
Most brand intelligence products are editorial. They watch a category, pick winners, and write a story. ShurIQ is different — it treats a market the way a regulator treats a banking sector: observe transactions, compare against a published rubric, rank, surface deviation, re-run on a cadence. The rubric is the law. The ontology is the evidence schema. The stack rank is the examiner's report.
We have established the rubric that establishes the ideal state for the different types of categories we rate. Then we map each brand and entity as they go about their business operations — observing the various events and flow of value, downloads, money spent, deals made, shows watched. — Operating brief, session 2026-04-15
The vertical drama build is the first full instance of this pattern. Nine minutes of user brief produced four parallel research slices, 261K tokens of intelligence gathering, one master ontology document, and — captured here — the knowledge graph that now becomes the long-lived substrate. Every future report against this sector queries the graph. The graph queries the rubric. The rubric returns a stack rank.
The Brand Power Score is a composite on [0, 100]. Each of five dimensions scores independently on the same scale; weights are set per vertical so that an entertainment app is measured differently than a nonprofit. Stack ranks are only valid within a vertical — the point of the rubric is to enforce comparability among peers, not across sectors.
Unaided recall, reach, share of voice. For vertical drama: MAU, app-store rank, organic search volume.
Paywall conversion, refund rate, store rating, regulatory posture. Legal disputes deduct.
Coherence of content slate, editorial stance, IP strategy, creator relations posture.
What the brand does that no peer does. For vertical drama: SAG-AFTRA status, exclusive IP pipelines, unit-economic moat.
ARPPU, 30-day retention, repeat-paywall crossings, net-promoter on forum cohorts.
Composite = 0.20·Awareness + 0.25·Trust + 0.15·Mission + 0.20·Differentiation + 0.20·Loyalty · weight vector: vertical-drama-apps-v1
An ontology is not drafted — it is composed from parallel investigations and then reconciled. The vertical drama ontology was built in eight observable phases over roughly thirty minutes of wall time, including fourteen minutes of simultaneous subagent research and a single forty-one-thousand-character synthesis write. The trace below is the audit trail for that build.
Build a vertical-drama industry ontology to enrich ShurIQ stack-ranking reports. No prior work to collide with. Target surface: competitive-intelligence folder.
Locate existing ShurIQ reports. Map the competitive-intelligence folder. Confirm the target subdirectory does not yet exist. Create it.
Four subagents, one per thematic slice — companies & financials, executives & deals, content & platforms, regulatory & legal-tech. Each armed with SerpAPI + Exa + Firecrawl keys and an output-file target.
Subagents report over a five-minute window. Each returns a summary, a top-5 insight list, and an output file path. Combined: 129 tool calls across workers, ~261K tokens of intelligence synthesis.
Main agent reads each source slice to verify structure and surface cross-slice patterns. Hits token ceiling twice; recovers by narrowing read windows.
Single large Write call creates the master ontology: exec summary, entity registry, deal ledger, genre ontology, regulatory stack, financial-signals index, ShurIQ integration proposal with 12 scoring dimensions and 5 bridge prompts.
Ontology promoted into InfraNodus as a live knowledge graph. 50 relationship statements encode entities, deals, regulators, content categories, and rubric dimensions. Graph ingests to 98 nodes across 16 clusters — modularity 0.828.
The graph is canonicalised as a source. Future reports will interface with it — querying entities, scoring against the rubric, producing stack ranks on cadence. This document defines the primitives by which those reports render.
Below is a curated constellation drawn from the seeded graph. Thirty-two of the ninety-eight nodes are rendered — the ones with highest betweenness centrality, one rubric dimension per cluster, and enough structural connective tissue to make the power structure legible. Each node's radius encodes its importance (square-root of degree, capped at 32 pixels). Cluster colour encodes community assignment. Edge thickness is uniform; opacity rides the constellation band [0.04, 0.15] to keep the page calm.
Source graph: infranodus.com/sensecollective/vertical-drama-ontology · 98 nodes · 142 edges · 16 clusters
Consolidation in this sector does not run through mergers. It runs through equity stakes, accelerator placements, and Singapore–China IP decoupling. The following flows are the observed control surfaces — the lines the ontology draws that an editorial read of the market would miss.
A ShurIQ report is not a deliverable — it is a snapshot of a loop. The loop runs on cadence (ninety-day default for vertical drama). Each pass observes new events, re-scores the rubric, and re-ranks. The graph is the substrate that makes the loop cheap: re-running is a SPARQL-style query, not a re-read of the world.
Downloads, spend, deals, shows watched, regulatory filings. Events land in the ontology with timestamp and source URL.
New statements are appended. Existing statements are versioned. The knowledge graph absorbs the delta; centrality metrics re-compute.
Each entity is projected against the rubric. Each dimension resolves to a [0, 100] score. Composite is weighted per vertical.
Within-vertical stack rank is emitted. Position deltas against prior run are flagged. Crossings of threshold become alerts.
Output is filed as the next canonical report. Cadence ticker advances. Ontology is left open for the next pass.
Every ShurIQ report composes from the same seven marks. When you see one of these primitives, you know what is being claimed — and what query produced it. A report that introduces a new primitive triggers a grammar extension review; an ad-hoc shape is a bug, not a feature.
A nameable observed thing: a company, exec, app, deal, regulator, or rubric dimension. Radius encodes degree; colour encodes cluster.
Directed arrow encoding a verb from the REA ontology — equity, licenses, signals, deconsolidates, pays, unlocks. Labelled on hover.
A dated occurrence: deal close, regulatory filing, product launch, lawsuit. Anchors an edge to a specific timestamp on the cadence.
A rubric dimension rendered as a vertical axis [0, 100]. A brand's score is a bar against the axis; peer median is a dashed line.
A structural hole between two clusters that should be connected but are not. Rendered as a dashed edge in critical. Reads as a research question.
A proposed entity or concept that would close a gap. Drawn in accent-blue as a third node connecting two clusters. The output of bridge-prompt generation.
A statement's position on the subjectiveness spectrum [0, 1] — personal → team → emerging → professional → legal. Controls the floor at which a fact is admissible to a given report.
The vertical drama ontology is now a source. The graph is live. The rubric is calibrated. The grammar is documented. The next move is to run the loop — score Tier 1 Asian apps against the rubric, emit the first canonical stack rank, and let the cadence begin.
Live graph: infranodus.com/sensecollective/vertical-drama-ontology · 98 nodes · 142 edges · modularity 0.828 · refresh cadence 90 days