Capability detailTelemetry capability layer

Data Team

From dashboard-first reporting to evidence-first telemetry validation. Data Team explains whether telemetry, field counts, classification rules, and claim boundaries are strong enough for public-safe insight.

Role type

Capability and validation-readiness layer

Primary shift

Dashboard-first reporting to evidence-first telemetry validation

Core output

Validation-readiness notes and claim-safe insight inputs

Boundary

No production telemetry verification by default

Why this role exists

Dashboards and reports can look convincing before the underlying data source, field counts, classification rules, and claim boundaries are strong enough. Data Team exists to move from dashboard-first reporting to evidence-first telemetry validation.

What was upskilled

Before

first-pass profiling
dashboard-shaped export review
category counts without enough field-level proof

After

telemetry semantics discipline
field-level data quality validation
reproducible classification rules
sample-row proof planning
observed / inferred / not-claimable separation
visualization readiness gates before graph implementation

Permanent skill pack

data-team-core-telemetry-semantics

Telemetry meaning, naming discipline, signal and field semantics, and claim-safe interpretation.

data-team-quality-validation

Field-level checks, count reconciliation, reproducible classification rules, and validation-readiness evidence.

data-team-profiling-monitoring

Profiling, drift and quality awareness, monitoring-readiness, and owner-readable insight preparation.

Operating principles

Data source before dashboardField counts before category claimsClassification rules before visualizationObserved / inferred / not-claimable separation before insightSample-row proof before strong conclusionVisualization readiness before graph implementationClaim boundary before public-facing interpretation

External references and how they are used

OpenTelemetry Semantic Conventions

Used as an adapted reference foundation for telemetry signal meaning, naming discipline, attributes, resources, logs, metrics, traces, and events.

Great Expectations / GX-style validation

Used as an adapted reference model for expectation-style checks, validation results, and documenting data quality.

Soda-style checks

Used as an adapted reference model for pipeline-friendly checks, data contracts, source validation discipline, and scan-style quality gates.

Evidently-style profiling and monitoring

Used as an adapted learning foundation for profiling reports, feature and statistic tracking, monitoring-readiness, and owner-readable visualization insight.

How this role applies the skill

Field inventory - field-level counts - classification rules - category recount - observed / inferred / not-claimable split - sample-row proof - visualization readiness - insight report.

Inputs / Outputs

Inputs

telemetry exports
field inventory
field-level counts
classification rules
sample-row proof plan
visualization readiness criteria

Outputs

validation-readiness notes
quality findings
classification boundary notes
insight report inputs
visualization readiness decision

Escalates to

Gate / Owner

claim boundary or public interpretation requires approval

Checker

validation evidence needs deterministic review

Surface / UI role

visualization readability or page treatment needs review

Engineering / Data implementation role

source validation or pipeline work is needed

Evidence

Available Evidence

Data Team role/capability detail exists on /org-roles/data-team.
Capability learning narrative is documented.
External references are mapped to how they are applied.

Missing Evidence

Production telemetry verification is not claimed.
Provider-backed telemetry verification is not claimed.
Full row-level completeness is not claimed.
Live monitoring is not claimed.
Production graph readiness is not claimed.

Claims Not Made

No production telemetry verification.
No provider-backed telemetry verification.
No full row-level completeness.
No live monitoring.
No production graph readiness.
No framework certification claim.

Upskill Path / Promotion Evidence

Level	Label	Evidence to promote
1	Field Inventory Operator	can identify fields and meaning boundaries
2	Data Quality Validator	can define field-level checks and reconcile counts
3	Telemetry Semantics Reviewer	can separate observed / inferred / not-claimable and protect claim boundaries
4	Visualization Readiness Lead	can decide when data is ready for graph or insight presentation