Spatial Stack with Matt Forrest – Details, episodes & analysis

In this episode of the Spatial Stack, Matt sits down with Kyle Walker, Professor of Geography at TCU and the creator of popular R packages like tigris and tidycensus.

Kyle dives into why he views US Census data as critical infrastructure and how open data is fundamentally transforming decision-making across industries like real estate and energy. He shares the origin story of his open-source work, explaining why he champions the R programming language for full-stack geospatial analysis. The conversation also covers the evolution of web mapping, from the laborious process of rendering dot-density maps to the blazing-fast performance of modern tools like PMTiles.

Finally, Kyle reveals how generative AI specifically Claude Code and the Zed editor is serving as his ultimate coding assistant, allowing him to rapidly build complex projects like the mapgl package and turn his ideas into reality faster than ever.

Connect with Kyle:

X/Twitter: https://x.com/kyle_e_walker
LinkedIn: https://www.linkedin.com/in/walkerke/
Bluesky: https://bsky.app/profile/kylewalker.bsky.social

00:01:00 – Welcome and Kyle Walker’s Background at TCU
00:06:18 – Why US Open Data is Critical Infrastructure
00:09:20 – Demystifying Census Data with tigris and tidycensus
00:15:48 – Applied Spatial Data: Real Estate and Forecasting Models
00:18:28 – The Evolution of High-Resolution Dot Density Maps
00:23:48 – The Human Element: How People React to Seeing Data Maps
00:29:14 – R vs. Python: Why R is a Geospatial Powerhouse
00:37:44 – Accelerating Development: Using Claude and AI for Coding
00:43:40 – The Future of Mapping: PMTiles, Segment Anything, and LLMs
00:48:18 – Where to Find Kyle’s Book, Tools, and Workshops

---

🚀 Join The Spatial Lab:
Stop guessing at your career path. Get direct mentorship, advanced training, and a roadmap to these high-value roles inside The Spatial Lab.
👉 https://forrest.nyc/spatial-lab/

📰 Daily modern GIS insights: https://forrest.nyc

CONNECT WITH ME
📸 Instagram: https://www.instagram.com/matt_forrest/
💼 LinkedIn: https://www.linkedin.com/in/mbforr/
📧 Newsletter: https://forrest.nyc
🌐 Website: https://forrest.nyc

We have never had more data about our planet: petabytes of satellite imagery, aerial photos, and sensor readings collected daily. Yet, turning that massive volume of "noise" into a clear signal remains the fundamental challenge of the geospatial industry.

In this episode of the Spatial Stack, I sit down with the engineering and product minds from Wherobots: Ryan, Phil, and Len - to tear down the architecture required to handle Earth Observation data at a planetary scale. We move beyond the buzzwords to discuss the engineering "war stories" of building resilient inference pipelines.

We dive deep into why the industry is moving away from simple computer vision toward "Large Earth Models" that function like LLMs for the physical world. We also get into the weeds of the tech stack: the battle between Dask and Ray for distributed compute, why Cloud-Optimized GeoTIFFs (COGs) aren't always the answer for inference, and how formats like Zarr are unlocking multidimensional analysis.

In this episode, we cover:

The Data Bottleneck: Why "garbage in, garbage out" is still the biggest hurdle in monitoring a changing planet.

Infrastructure Realities: The specific limitations of Google Earth Engine and why we needed a cloud-agnostic approach.

Engineering Pivot: Why Wherobots migrated from Dask to Ray to solve "crashing cluster" syndromes and memory management issues.

The Future of GeoAI: How embeddings and foundation models are compressing petabytes of data into searchable, semantic insights.

✅ Sign Up for Wherobots: https://wherobots.com/
✅ Learn more about Apache Sedona: https://wherobots.com/apache-sedona/
✅ Learn more about RasterFlow: https://wherobots.com/blog/rasterflow-earth-observation-inference-engine/
✅ Sign Up for the RasterFlow Private Preview: https://wherobots.com/rasterflow-preview/

00:00 – Teaser: The "Garbage In, Garbage Out" problem in GeoAI
00:01:51 – Introductions & Icebreakers (The controversial ice cream opinions)
00:03:08 – The Challenge: Monitoring a changing Earth at scale
00:10:30 – Data Engineering: The hidden complexity of NAIP, clouds, and tiling artifacts
00:14:19 – Modeling Reality: Why Computer Vision models fail on geospatial data
00:21:51 – The Google Earth Engine Debate: Walled gardens vs. bringing compute to the data
00:27:53 – Introducing Rasterflow: A new architecture for scalable inference
00:36:51 – The Engineering Story: Why we switched from Dask to Ray
00:43:40 – File Formats: Why Zarr is superior to COGs for multidimensional inference
00:47:40 – Workflow Walkthrough: Running the "Fields of the World" model
00:51:40 – Embeddings, Foundation Models, and Large Earth Models
00:57:40 – How to get started with Rasterflow

📰 Modern GIS insights: https://forrest.nyc

CONNECT WITH ME
📸 Instagram: https://www.instagram.com/matt_forrest/
💼 LinkedIn: https://www.linkedin.com/in/mbforr/

🌐 Website: https://forrest.nyc

📩 Get every update from my newsletter ➡️ https://forrest.nyc ⬅️

What if you could stop wrangling geospatial data, and just get to the good part?

In this episode, I sit down with Rob Fletcher, Chief Science Officer at Seer AI, to talk about one of the biggest blockers in geospatial today: accessing and integrating the right data, at the right time, in the right format.

Rob shares his journey from doing particle physics at CERN to pioneering spatial data infrastructure, and how Seer AI is building what he calls a “CDN for geospatial data.”

We dive into:

🌍 Why the biggest challenge in GIS isn’t analysis—it's finding and formatting data
💡 The danger of agreeing to analytics projects without knowing what's possible
🛠️ Why Seer AI chose graph technology to organize spatial data across sources
⚡ How a "side project" dashboard became essential to real-time disaster response
📊 Bridging the gap between GIS teams and business analysts

Whether you're working in local government, disaster response, sustainability, or just tired of managing messy spatial data, this episode will change how you think about building modern GIS workflows.

🔗 Learn more about Seer AI: https://seer.ai
🎧 Subscribe for more episodes on the future of geospatial

📩 Get every update from my newsletter ➡️https://forrest.nyc ⬅️

🔑 Sign up for Placekey here:https://placekey.io

Merging and cleaning location data is one of the biggest headaches in GIS, real estate, and business intelligence. Duplicate addresses, inconsistent formats, and messy datasets can derail even the most sophisticated analysis. But what if there was a universal identifier that made all of those problems disappear?In this episode, I sit down with Hayden Mortimer, President of PlaceKey, to uncover why traditional addresses aren’t enough and how PlaceKey is revolutionizing the way we connect place-based data. We dive into the challenges of entity resolution, the pitfalls of fuzzy matching, and how a simple, open-source identifier is making spatial joins seamless.Whether you work in GIS, data science, or any field that relies on location data, this conversation will change how you think about addressing and spatial relationships. Don’t miss it!✔️ Why addresses alone aren’t reliable for data integration✔️ The hidden challenges of matching places across datasets✔️ How PlaceKey works and why it’s different from other solutions✔️ Real-world use cases for small businesses and enterprise-level applications✔️ The future of location data and entity resolution🎧 Listen now and learn how to fix your messy location data for good!

In this conversation with Kyle Barron from Development Seed, we will learn lonboard, a new Python package that solves the problem of visualizing massive geospatial datasets in Python. Learn about the motivations behind the project, some of the tech under the hood, and of course a demo!

Sign up for TLDR for free here!

Meta has more than 3 billion users across Instagram, WhatsApp, and even its new AR glasses. Behind the scenes, all of them are powered by one thing: maps. But instead of relying on closed systems, Meta is betting big on open data—and building its own global map.

In this episode, I talk with Said Turksever from Meta, who leads their open mapping strategy. We dive into:

🌍 Why Meta cares so much about maps
🛠 The tools they’re building with AI and open source
🏙 How cities from Phoenix to Naples are being transformed by open data
🚶 The future of pedestrian mapping and accessibility
🤝 The role of communities in shaping the next generation of maps

From disaster response to daily navigation, the impact of open mapping stretches far beyond social media. This is a conversation about technology, community, and the future of how we navigate the world.

🚀 Ready to move beyond desktop GIS?
Step into the Spatial Lab: a global community for ambitious geospatial professionals who want to break out of outdated workflows and join the top 5% of the field.
👉 Join Spatial Lab: https://forrest.nyc/spatial-lab/

🎓 Want structured, career-changing learning?

🚀 Modern GIS Accelerator: https://forrest.nyc/accelerator/
— master Python, Spatial SQL & cloud workflows in 6 weeks

🧭 Career Compass: https://forrest.nyc/career-compass/
— fast, practical steps to land the GIS role you want

🪄 AI Copilot for GIS: https://forrest.nyc/ai-copilot-for-gis/
— learn to integrate AI into your geospatial workflows & boost your productivity

📰 Weekly modern GIS insights: https://forrest.nyc

⚡️ Spots for the next live cohort and mentorship cycle are closing soon,  join now to lock in your place and momentum.

CONNECT WITH ME
📸 Instagram:  https://www.instagram.com/matt_forrest/
💼 LinkedIn: https://www.linkedin.com/in/mbforr/
📧 Newsletter: https://forrest.nyc
🌐 Website: https://forrest.nyc

📩 Get every update from my newsletter ➡️ https://forrest.nyc ⬅️
💨 Sign up for the Spatial Lab Community ➡️ https://forrest.nyc/spatial-lab ⬅️

What if the future of AI depends on understanding where, not just what? In this episode, I’m joined by Justus Killian, partner at Space Capital, one of the only VC firms focused entirely on geospatial and space-based tech. We dive into why geospatial data is no longer niche, how satellites became billion-dollar businesses, and why the next wave of AI breakthroughs could come from those who understand location. We also unpack the evolving venture landscape, the rise of geo-embeddings, and how spatial data is finally moving beyond the map into real-time decision-making.

Whether you're building LLMs, funding the next Planet Labs, or just curious about how Earth is being mapped in ways you’ve never seen, this one's for you.

Join Isaac Brodsky, CTO and Co-Founder at Fused where we will dive deep into scaling geospatial workflows and shipping them using their new product Fused.

https://www.fused.io/

Matt returns to discuss the future of Spatial Stack—what it is, where it’s headed, and how it will evolve into a go-to platform for exploring the intersection of geospatial technology and real-world impact. From its origins as a live-stream experiment to a full-fledged video-first podcast, this episode lays out the vision for breaking the geospatial echo chamber, identifying new opportunities, and working together to expand the field.

Expect a mix of solo deep dives, expert conversations, and practical discussions on geospatial trends, Earth observation, and the growing role of spatial data across industries. Whether you're a seasoned GIS professional or just curious about geospatial’s future, this episode sets the stage for what’s to come.

How do you get access to just the data you need? And what if that data isn't in large datasets or traditional mapping services? And can AI make this data accessible to those without a background in geospatial? Get answers to these questions and more in this live stream with Hantz Févry, CEO of Stoovo!