Oh the places you’ll go with spatial data

Summary of Oh the places you’ll go with spatial data

by The Stack Overflow Podcast

29mJune 23, 2026

Overview of Oh the places you’ll go with spatial data

This episode of the Stack Overflow Podcast explores how map data is built, standardized, versioned, and shared at scale, with a focus on the collaboration between Microsoft Maps and the Overture Maps Foundation. Host Ryan Donovan speaks with Jeffrey Hightower (VP of Places Data at Microsoft) and Amy Rose (CTO of Overture Maps Foundation) about why open, interoperable spatial data matters, how the foundation works, and what it means for developers and companies building location-based products.

What map data actually is

The guests break down map data as more than just points on a map:

  • Spatial data can be:
    • a single coordinate
    • a 3D point with elevation
    • a boundary or polygon
    • a transportation segment like a road or rail line
  • It often includes attributes/features such as:
    • road direction or one-way status
    • building height
    • business category
    • whether an area is forested or a park

They also explain that map data is collected in many ways, including:

  • field collection
  • measurements
  • user contributions
  • satellite/image analysis
  • data aggregation from multiple sources

Why Overture Maps Foundation exists

Amy Rose explains that Overture Maps Foundation was founded in late 2022 to reduce duplicated effort across organizations that were all building their own foundational map datasets.

Core mission

  • Create global spatial datasets that are:
    • open
    • standardized
    • interoperable
  • Let organizations focus on their differentiating products, rather than rebuilding the same map infrastructure

Founding members and growth

  • Founded by:
    • Microsoft
    • Meta
    • AWS
    • TomTom
  • Esri joined soon after
  • The foundation has grown to nearly 50 member organizations across:
    • academic
    • nonprofit
    • public
    • private sectors

Standardization, schemas, and GeoParquet

A major topic in the conversation is how organizations agree on the structure of map data.

Key technical choices

  • Overture aligned on GeoParquet early as a distribution format
  • The group also collaborated on the schema for representing:
    • roads
    • parks
    • buildings
    • places/businesses

Why this matters

  • Better alignment means less wasted effort
  • More perspectives lead to stronger technical decisions
  • Open collaboration helps avoid building closed systems that only fit one company’s needs

Amy notes that even though schema discussions can be contentious, that friction is part of making a better shared system.

GERS: the global entity reference system

One of the most important concepts discussed is Overture’s Global Entity Reference System (GERS).

What GERS does

  • Assigns a stable, persistent ID to each entity
  • Tracks things like:
    • buildings
    • road segments
    • addresses
    • places/businesses
  • Preserves identity across releases, even as attributes change

Why separate IDs matter

Buildings, addresses, and businesses can change independently, so Overture tracks them separately to keep the data more accurate and interoperable.

Practical benefit

This makes it possible to attach other datasets over time, such as:

  • business ownership
  • restaurant menus
  • proprietary business data
  • licensing data from other providers

Versioning and release management

The episode spends a lot of time on how you version a world that is constantly changing.

Overture’s versioning model

  • Monthly releases for themes like:
    • transportation
    • buildings
    • places
  • Each release has a schema version
  • Each individual record also has its own feature version

What users can track

With GERS and the changelog, users can see:

  • when a feature first appeared
  • the most recent release it appeared in
  • how many versions it has gone through
  • what changed over time

Why it matters

This supports:

  • backward compatibility
  • data lineage
  • reproducibility
  • easier developer workflows

Open access for developers

The foundation is designed to be easy to use.

How developers can access it

  • Data is available in public AWS and Azure buckets
  • No authentication required
  • Ways to explore the data:
    • browser-based Overture Maps Explorer
    • Python CLI/library
    • direct SQL querying with DuckDB

Who it is for

  • Primarily developers
  • Especially teams building:
    • map products
    • search systems
    • location-aware applications
    • AI systems that need grounded geographic data

Jeff notes that the data is meant for builders, not casual consumers.

How Microsoft uses Overture data

Jeff explains that Microsoft uses Overture as a shared foundation while still keeping its own differentiated systems.

Microsoft’s approach

  • Ingest Overture data into internal pipelines
  • Combine it with Microsoft’s own proprietary data
  • Focus internal investment on areas that matter most to its business, such as:
    • local search
    • Bing
    • Copilot
    • web crawling and indexing

Shared infrastructure

Microsoft also reuses open-source tooling from Overture, including:

  • Spark
  • Apache Iceberg

This reduces duplicated work and lowers the cost of maintaining map infrastructure.

Data licensing and responsible use

The hosts briefly discuss whether map data could be used for less desirable purposes, such as building mailing lists.

Overture’s stance

  • Uses highly permissive licenses where possible
  • Common licenses include:
    • CC0
    • CC BY 4.0
  • Source and licensing are recorded at the record level
  • Users are expected to follow the applicable terms of use

Surprising use cases and community impact

Amy shares one of the most interesting examples she has seen: using Overture places data to build a global map of gluten-free restaurants.

This illustrates one of the big benefits of open spatial data:

  • unexpected applications
  • community experimentation
  • rapid innovation
  • feedback that improves the dataset over time

What’s next for Overture Maps

The conversation closes with a look at the future.

Likely near-term direction

  • Continue the current release model
  • Improve how data is exposed to support AI and grounding
  • Make upstream artifacts more available, such as:
    • metadata
    • validation results
    • provenance information
    • source-matching details
    • freshness signals

Longer-term frontier

  • Better support for fresh/live data
  • More support for live ingest
  • More community-driven innovation around real-time spatial updates

Key takeaways

  • Map data is a complex, layered dataset with geometry, attributes, and provenance.
  • Overture Maps Foundation exists to make foundational map data open and interoperable.
  • GERS IDs are central to keeping entities stable across releases.
  • Overture’s value is not just the data, but also the schema, versioning, and processing pipelines.
  • Microsoft benefits by reducing redundant work and focusing on differentiating products like search and Copilot.
  • The open model enables both practical developer use and unexpected community-driven applications.