What does it really take to teach AI to understand our planet?

In this episode, Matt sits down with Isaac Corley, Senior Machine Learning Engineer at Wherobots and maintainer of TorchGeo, and Christopher Ren, Data Scientist and writer behind some of the most thought-provoking essays on Earth observation and AI.

They dive deep into the state of Geospatial AI from embeddings and foundation models to how these tools are reshaping Earth observation and remote sensing. You’ll hear real-world perspectives on what’s working, what’s not, and where the hype outpaces reality.

Key topics covered:

What "embeddings" actually mean for geospatial and remote sensing
How foundation models like AlphaEarth and Tessera are trained
The challenges of applying ML at global scale
The future of Earth observation data pipelines
Why geospatial AI is still the "wild west"

If you’ve ever wondered how AI models are built to map, monitor, and understand the Earth, this conversation breaks it down without the buzzwords.

LINKS:

ISAAC CORLEY

Website: https://isaacc.dev/
LinkedIn: linkedin.com/in/isaaccorley
X: @isaaccorley_
GitHub: @isaaccorley
I’m active in TorchGeo slack: free to join https://join.slack.com/t/torchgeo/shared_invite/zt-22rse667m-eqtCeNW0yI000Tl4B~2PIw

CHRISTOPHER REN

Blog: https://christopherren.substack.com/
Podcast: Spotify | Apple | Youtube
LinkedIn: https://www.linkedin.com/in/christopherren/
Geovibes/Github: https://github.com/cr458/geovibes

🚀 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Step into the Spatial Lab: a global community for ambitious geospatial professionals who want to break out of outdated workflows and join the top 5% of the field.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 2 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

AI can answer almost anything, except “where.” That gap is huge, and  Kieran McCann from Mapbox is working to close it.

In this episode of The Spatial Stack, Matt sits down with Kieran, who’s spent six years helping Mapbox quietly power some of the most widely used maps and location services in the world. From his early days doing grunt map edits at Apple Maps to shaping Mapbox’s next big bet: AI-native mapping. Kieran shares what it takes to bring true spatial understanding into artificial intelligence.

You’ll hear how Mapbox is experimenting with ways for AI to finally reason about place and direction, what’s still missing for models to “get” geography, and how anyone, even those just starting out, can build powerful AI + geospatial tools today. Kieran also breaks down how showing prototypes can change your career trajectory and why now is the perfect time to jump into this space.

If you’re curious about where mapping and AI collide and what that means for your work, your career, and the apps we all use every day, this conversation is a must-listen.

🚀 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 2 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

GeoAI isn’t just another buzzword, it’s a turning point for the entire geospatial industry. In this episode of The Spatial Stack, Matt sits down with Tee Barr, Director at Verisk and Councilor for the American Geographical Society, to unpack why GeoAI could reshape the field as profoundly as GPS once did.

From climate risk and supply chain resilience to defense, finance, and human security, Tee explains how AI is already influencing high-stakes decision making and why the upcoming Geography 2050 event is dedicating its theme to this massive shift.

You’ll hear how GeoAI is challenging long-held assumptions about spatial data, the new opportunities and risks it creates, and why ethical responsibility must keep pace with technical innovation. Whether you’re a geospatial professional, business leader, or educator, this conversation will help you understand the stakes and the opportunities in the next era of mapping and analytics.

Register for Geography 2050 here: https://www.geography2050.org/

---

🚀 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 2 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

What happens when a geospatial professional decides to map the entire business of our industry?

In this episode of Spatial Stack, Matt talks with Wilfred Waters, the creator of the first-ever Geospatial Index, a hand-built list tracking hundreds of public companies shaping maps, data, satellites, sensors, and spatial software. Will shares how he spent countless late nights pulling together every geospatial stock he could find, from Vietnam to Finland to the U.S., and what the process revealed about where the industry is heading.

Along the way, we unpack how indexes work, why understanding the business side of geospatial matters (even if you’re not an investor), and how seeing the market clearly can help you think differently about your own career and the technologies you rely on.

Geospatial FM on Spotify: https://open.spotify.com/show/5gpQUsaWxEBpYCnypEdHFC
On Apple Podcasts: https://podcasts.apple.com/us/podcast/geospatial-fm/id1674368928
Substack: www.geospatial.fm
Index introduction and construction rules: https://www.geospatial.fm/p/compounding
2024 Year In Review: https://www.geospatial.fm/p/2024-geospatial-stocks-performance

⚠️ Nothing in this episode is investment advice. It’s for educational and informational purposes only.

What if satellites became as smart and as accessible as your smartphone?

In this episode, I sit down with Michael Bartholomeusz, CEO of NOVI to unpack how Earth observation is being transformed by edge computing in space. We explore the shift from sending terabytes of raw imagery to delivering actionable insights, and why this change could take the cost of analysis from $150,000 down to under $1,000.

Michael shares stories from his 25-year career, from working on the space shuttle that launched John Glenn, to leading NOVI’s dual-use strategy for defense and commercial markets. Along the way, we dive into:

Why the current Earth observation model is broken
How edge computing flips the economics of space data
The “iPhone moment” for satellites and what it means for new applications
Real-world examples of cost savings and new use cases
The industries poised to benefit most — from defense to agriculture to disaster response

This is a conversation about more than satellites. It’s about making Earth observation affordable, democratized, and insight-driven.

NOVI’s Website: www.novispace.ai
NOVI’s LinkedIn: https://www.linkedin.com/company/novispace-inc
NOVI’s GENIE™ Constellation Press Release: https://www.newswire.com/news/novi-unveils-genie-tm-a-platform-and-satellite-constellation-that-22622479

In this episode, I sit down with Dr. Qiusheng Wu, associate professor, open-source advocate, and creator of some of the most widely used tools in modern geospatial, including Leafmap and GeoAI. If you’ve ever watched his YouTube tutorials or come across his projects on GitHub, you already know the impact he’s had on the community.

We dive into:

His newly released Python for Geospatial book, one of the most complete guides in the field.

The ongoing debate of Python vs. SQL in modern workflows—and how to combine them effectively.

The rise of embeddings and geospatial foundation models (including Google’s AlphaEarth) and what they mean for GIS.

The future of open-source education and why anyone, with patience and time, can learn and contribute.

How Dr. Wu balances teaching, tool-building, publishing, and pushing the boundaries of what’s possible.

This is a wide-ranging conversation about technology, education, and the future of spatial data. If you care about where geospatial is headed—and how to keep up—you won’t want to miss it.

👉 Links

Dr. Wu's Book on Geospatial Python: https://a.co/d/7M08rlY
Website: https://wetlands.io/
Dr. Wu on YouTube:  @giswqs  
LinkedIn: https://www.linkedin.com/in/giswqs/

Open data is supposed to unlock innovation, but what happens when “open” isn’t really open?

In this episode of The Spatial Stack, I sit down with return guest Mina Nada to explore the world of open data licenses — from Creative Commons to Open Database Licenses and beyond. We break down what these licenses actually allow you to do, where the pitfalls are, and how they impact anyone building geospatial products, tools, or services.

Along the way, Mina shares real-world experiences navigating licensing challenges, from NASA’s permissive CC0 approach to the more restrictive “share-alike” models. Whether you’re a developer, analyst, or entrepreneur, this conversation will help you understand how to work with open data without running into legal or business roadblocks.

If you’ve ever wondered what “open data” really means for your projects, this is the episode to listen to.

The future is spatial — so let’s get into it.

What happens when a hackathon project becomes one of the most influential open-source BI tools in the world?

In this episode, we sit down with Max Beauchemin — creator of Apache Superset and Apache Airflow — and Evan Rusackas from Preset, the company behind Superset’s enterprise evolution.

We explore:

The origin story of Superset at Airbnb — and why Max says it became his “life’s mission”
How open-source BI tools are outpacing traditional vendors like Tableau and Power BI
Why geospatial data is more than just maps — it’s a core analytical dimension
The design philosophy behind Superset: from SQL-first thinking to visual storytelling
Lessons from building geospatial capabilities into dashboards with Deck.GL and H3
The future of BI, and why the open ecosystem is winning

Whether you're a data engineer, GIS pro, or just curious how modern BI is evolving, this episode is packed with insights from two people who are building that future.

📍 Guest Links:

Apache Superset: https://superset.apache.org/
Preset: https://preset.io/

🎧 Tune in now and hear how Superset is redefining what modern, open, and geospatial-aware BI looks like.

In this episode, I open up about something many of us in GIS rarely talk about — imposter syndrome. That nagging feeling of not being good enough, of questioning whether you really belong in this field, even when the evidence says otherwise.

This isn’t a polished success story, it’s an honest conversation about doubt, comparison, and the pressure to measure up in GIS. I share my own experiences, what I’ve learned along the way, and why so many professionals quietly wrestle with the same thoughts.

If you’ve ever wondered “Am I really cut out for this?” — you’re not alone. Let’s talk about it.

If you’re working with spatial data, AI workflows, or massive batch jobs, you’ve probably hacked together more than a few pipelines. But what if there’s a better way?

In this episode, I sit down with Kenten Danas, Senior Manager of Developer Relations at ‪@Astronomer‬ to explore how Apache Airflow powers the modern data stack including real-world geospatial and climate risk modeling pipelines.

We cover:

What Airflow actually is (and why it’s everywhere)
How it’s used in geospatial pipelines, AI, and LLM workflows
New features in Airflow 3.0 like assets, remote execution, and backfills
Why orchestration is the key to scalable spatial data processing
Tools like the Airflow AI SDK that make LLM pipelines easier to manage

Links from the show:

Astronomer Academy (with courses + certifications): https://academy.astronomer.io/
Astronomer Webinars: https://www.astronomer.io/events/webi...
Astro CLI (for running Airflow locally): https://www.astronomer.io/docs/astro/...
Free trial of Astro: https://www.astronomer.io/lp/signup/
Airflow AI SDK (open source Python SDK for working with LLMs from Airflow): https://github.com/astronomer/airflow...
Vibrant Planet Geospatial + ML Airflow use case: https://www.astronomer.io/blog/airflo...

Whether you're building spatial features for machine learning or just want a more reliable way to manage your data workflows this is the episode for you.