Pineapple on pizza

No data available

Feature ownership
  • Data models
    Views, materialized views, orchestration, and lineage.
  • Data quality checks
  • Endpoints
  • Semantic layer and data catalog
  • Warehouse properties
    Syncs tables and materialized views to person or group properties.
Members

Roadmap

Recently shipped

Clearer view failure digests

Materialized-view failure emails now separate views that need action from newly failing and ongoing failures. Each row includes the failure streak and, when relevant, what to do next.

Goals

Q3 2026 objectives

Objective 1: Semantic layer (owner: Thiago)

Motivation: We are increasingly hearing about inconsistency issues when an agent writes a query or an insight based on its assumptions around a definition. Same problem of "what's the source of truth?" like we humans have, but agent-amplified.

What we'll ship:

  • A central semantic layer for all metrics and definitions parts of PostHog that need to do analysis or generate queries on top of a user's shared business understanding

Objective 2: Integration with the warehouse (owner: Jovan)

Motivation: We need to move workloads away from ClickHouse and onto the warehouse — modeling and endpoints included.

What we'll ship:

  • Endpoints execution running solely on the warehouse with appropriate compute control
  • Move materialization jobs to the warehouse, and limit usage on CH
  • Support pricing and billing instrumentation

Objective 3: Version control (owner: Jovan)

Motivation: While data modeling can happen inside PostHog, lack of version control makes it difficult to collaborate and safely make changes.

What we'll ship:

  • An easy way for agents and humans to interact with their git-based data modeling project
  • And enable a development workflow that doesn't impact the production models

Objective 4: Testing (owner: Thiago)

Motivation: Our agent's SQL writing has significantly improved, but we still have the verifiability gap where the users do not have a way of checking whether the models they build act as intended.

What we'll ship:

  • Schema and data tests for data models
  • That are integrated into the version control workflow
  • And easily used by agents

Side quest 1 (owner: Anna)

Motivation: Not every customer has custom and complex data modeling needs. Some are very standard and can be templated into models that take a customer from 0 to 1 in modeling their revenue / active users / attribution.

What we'll ship:

  • A set of skills that provide all the context to an agent around what data models it needs to create in PostHog and how to verify them

Side quest 2 (owner: Anna)

Motivation: We need to stay on top of the data space and how it's evolving. A lot has happened recently, and we want to make sure that what we're building is aligned with the field.

What we'll do:

  • Introduce recurring data space wrap-up research automation

Handbook

Features

Slack channel

#team-data-modeling