Metrics and Semantics: An evolution in data consumption

on YouTube ↗40m

" summary

MetricFlow is an open source metric store that compiles YAML metric definitions into rendered SQL for Snowflake, BigQuery, Redshift, Databricks, Postgres, MySQL and DuckDB, validating configs and showing query plans; the team open sources after ~2.5 years of work and added Databricks in two days. It supports cumulative metrics, caching and lineage integration with dbt but not data cleansing, struggles with pre‑denormalized wide tables, and is working on conversion metrics and percentile support.

AI-written summary · about

" chapters (10)

1. 0:00Introduction
2. 1:48Speaker background
3. 3:57Analysts and data access
4. 5:21Reality to data narratives
5. 7:54Semantic layers and Metric Flow
6. 16:30Designing semantic layers
7. 21:10Benefits and challenges
8. 25:02Metrics central to semantics
9. 29:19Empowering analytical users
10. 30:22Q&A

" related talks

• 27mMetric StoreBen Hababo, Mickey Rozen · hayaData 2024
• 26mMetrics-driven Data ArchitecturesNed Letcher · DataEngBytes 2021
• 18mIntroducing a SQL-based metrics layer powered byMike Driscoll · DuckCon #6
• 28mStop querying tables: give AI governed metrics insteadNishant Bangarwa · The Fifth Elephant 2025 Winter Edition
• 54mThe Future of Data Science/ML at Scale: A Look at MLflow, Delta Lake + LakehouseJules S. Damji · Data Love 2021
• 35mBuilding a Real Time Metrics DatabaseJoel Barciauskas · Dash 2019
• 20mBuilding metric platform using Flink for massive scale at NetflixAbhay Amin · Flink Forward Virtual 2020
• 24mDataEngBytes 2023 - MEL-T3-03 - Akira Wang & Abdi FarahAkira Wang, Abdi Farah · DataEngBytes Melbourne 2023