AGENT SKILL

bi-semantic-layer-metric-trees

Define normalized metric logic in headless semantic layers (Cube, dbt Semantic Layer, MetricFlow) and decompose North Star metrics into hierarchical driver trees.

License 许可
MIT
最近更新
2026-11-09
Stars
★ 0
Forks
0

01 前言 Front Matter

name
bi-semantic-layer-metric-trees
risk
low
tags
[ "business-intelligence", "semantic-layer", "dbt", "metric-trees", "data-governance" ]
author
AAS Platform
source
authoring
version
1.0.0
category
business-intelligence
description
Define normalized metric logic in headless semantic layers (Cube, dbt Semantic Layer, MetricFlow) and decompose North Star metrics into hierarchical driver trees.
disable-model-invocation
false

03 SKILL.md 正文

Metric Trees and Headless Semantic Layer Engineering

Overview & Core Principles

Define normalized metric logic in headless semantic layers (Cube, dbt Semantic Layer, MetricFlow) and decompose North Star metrics into hierarchical driver trees.

Metric Tree Architecture & YAML Specification

  1. Mathematical Decomposition:
    • Top Level: ARR = Total Active Subscriptions * ARPU.
    • Tier 2 Drivers: ARPU = Base Tier Price + Expansion Revenue per Account.
    • Churn Metric: Net Revenue Retention (NRR) = (Beginning ARR + Expansion - Downgrades - Churn) / Beginning ARR.
  2. MetricFlow Semantic Manifest:
semantic_models:
  - name: orders
    model: ref('fct_orders')
    dimensions:
      - name: ordered_at
        type: time
        type_params:
          time_granularity: day
      - name: customer_country
        type: categorical
    measures:
      - name: gross_order_value
        agg: sum
        expr: line_total_cents / 100.0

metrics:
  - name: revenue
    type: simple
    type_params:
      measure: gross_order_value
  1. Consistency Verification:
    • Metric definitions are source-controlled in Git, tested with CI assertions, and queried uniformly by BI tools and reverse-ETL.

04 同仓库更多 Skills