Our methodology
How we turn data
into evidence.
A number is useful when you can trace it. Here is where our data comes from, how we define it, and what the evidence can actually support.
Sources, definitions and limitations stay with every chart.
01 / Sources
Start with the source.
We distinguish records observed on-chain from values reported by a provider and estimates derived from them. Each needs its own context.
On-chain records
Saved transaction and execution records underpin the observed conversion and treasury-payout cohort. Block data provides recorded supply boundaries.
Keep in mind. The conversion and payout study covers 6 August–3 November 2026 UTC and a selected treasury-linked cohort, not all market activity.
Reported fees & volume
Daily fee and volume series are retained with their provider's definition and observation dates. Source disagreements remain visible rather than being averaged away.
Keep in mind. These sources share Dune lineage. Agreement is not independent confirmation, and provider-reported gross fees are not attributable protocol revenue.
Prices & supporting context
Historical price observations support an estimated dollar valuation. Governance materials and indexed records provide additional, separately qualified context.
Keep in mind. A daily price is not a swap execution price. An index's observed history is not proof of every action, and a policy statement is not a guaranteed outcome.
Coverage, access and reuse rights vary by source. Historical observations keep their dates; a fresh retrieval does not make an old observation fresh.
02 / Definitions
Different numbers. Different meanings.
Activity, fees, purchases and supply answer different questions. We define a metric once and keep its meaning attached to every view.
| Measure | What we count | What it does not establish |
|---|---|---|
| Observed NEAR purchases | Successful conversions of other assets into wrap.near within the observed treasury-linked cohort. | Not whole-market demand or proof that earned fees funded the purchases. |
| Treasury payouts | Matched native transfers from intents.near to the three selected wallets, excluding internal treasury hops. | Cash flows, not revenue, profit or an independently additive measure of token demand. |
| Provider-reported gross fees | The provider's gross protocol and distribution fee measure, retained under its stated methodology. | Not revenue retained by NEAR. Definitions and historical values can differ or be revised by the provider. |
Treasury purchases do not reduce total supply — burns do.
Token amounts retain exact arithmetic. Dollar valuations based on daily prices are estimates. Missing observations remain gaps, never zeros; flows and prices are never silently filled forward.
03 / Preparation
A traceable path to every chart.
Validation comes before publication. The source, calculation and dataset version remain linked as a number moves into a chart or a research document.
Preserve the source
Keep the saved input, its source, observation period and content hash. Recorded evidence stays tied to its original dates.
Validate the records
Check identities, duplicates, successful execution, units and calendar coverage. Missing or conflicting records carry an explicit reason.
Apply one definition
Calculate each metric with a registered formula and exact token arithmetic. A changed definition receives a new version.
Publish a fixed version
Only validated datasets are approved. Charts identify the resolved version; saved research revisions and exports retain the evidence they pinned.
For example: observed NEAR purchases
Saved successful conversions in the treasury-linked cohort → deduplicated events and exact NEAR amounts → the versioned purchase metric → an approved dataset and its chart. The 90-day window is 2026-08-06 ≤ UTC date < 2026-11-04.
04 / Verification
Inspect the evidence behind a number.
Follow a number from its definition to its saved context before using it in your thesis.
Read the definition
Open Sources & definitions on a chart. Check what is counted, the unit, the source and the limits of the measure.
Check the selected evidence
Confirm the dates, filters and dataset version. A historical range remains historical even when retrieved today.
Inspect the exact values
Use the equivalent data table, available without hovering. Missing values remain gaps, and estimates or bounds retain their labels.
Keep the same context
Save the inspected selection to research. Its acknowledged revision and exported figures retain that dataset and calculation context.
05 / Interpretation
Know where the evidence stops.
A measured relationship is not a causal explanation. Coverage, accounting scope and the way a provider constructs its data limit the conclusions we can draw.
A cohort is not the whole market.
Observed purchases cover treasury-linked collector accounts only. Short attribution cohorts stay descriptive; the 90-day window cannot establish a long-term stable mechanism.
A bound is not an exact total.
Gross issuance and burn remain conservative lower bounds: the mint-header audit verified 6 of 297 headers. The unverified remainder is never assumed zero.
Correlation is not causation.
Fees and volume can be mechanically related. Shared sources, market trends and accounting choices remain limitations even when a statistical model reports a strong association.
Current model evidence
No reviewed fitted result is published here. The current data-readiness decision admits a fee-versus-volume proxy only with limits; it does not establish attributable protocol revenue or token value capture.
View published model evidence06 / Reproducibility
The same evidence, again.
A reproducible result needs more than a screenshot. Its inputs, definitions, selections and calculation environment must be identifiable.
Source & input identity
Where the observations came from, their actual period and the content hashes of the retained inputs.
Definition & selection
The metric and chart versions, units, dates and filters used for the figure.
Dataset & calculation
The approved immutable dataset, formula version and recorded transformation path.
A fixed research revision
The saved context behind an exported figure, rather than a chart that silently refreshes later.
Raw archive access remains restricted by the source and sharing permissions. No raw-data downloads are offered here. Published metric, chart and model views expose their permitted evidence and limitations.