How Vista is moving away from centralized data for good
summary
Vista is fully decommissioning its central warehouse and replatforming into a committed data mesh, rebuilding ten years of work into a Snowflake-backed mesh with Databricks, dbt and Airflow. The org creates 25 data product teams and ships ~100 data products (dashboards, tables, APIs), including a self-service Lines of Business Explorer with 630 fields and a personalization industry classifier. Learnings focus on a data domain map, strong platform tooling, product ownership, and unresolved pains: cross-team interoperability, decentralized data modeling and sensible standards.
AI-written summary · about
chapters (14)
- 0:00Intro
- 1:02Title
- 1:19What is Vista
- 2:28About Vista
- 2:48Data and analytics
- 3:28Data like water
- 4:28Old data stack
- 12:20Data Mesh advantages
- 14:25The four pillars
- 14:42Strong foundations
- 19:08Fastpaced innovation
- 21:26Product thinking for data
- 23:29Interoperatability
- 28:17Outlook
related talks
36mIntroducing Data Mesh A Paradigm Shift In Analytical Data Management Architecture
31mData Mesh & Data Vault
48mData Mesh Paradigm Shift in Data Platform Architecture -
39mKeynote - Data Mesh
19mKickstarting a greenfield data project
39mData Mesh: Why, What and How
41mOn Data: Sky above, sand below, peace within
44mData Mesh: Lessons from the trenches