Podcast
Geospatial FM
Hosted by Wilfred Waters · EN
Obsessed with geospatial foundation models, broadcasting in geospatial and compounding via publicly listed geospatial equities.
14episodes
Episodes
Newest firstAll episodes
Listening for Whales with NV5 Geospatial
6d ago00:43:25Tap to summarizeMark Baumgartner is a senior scientist at Woods Hole Oceanographic Institution. Back in 2011 he and Sarah Mussoline published a paper on how to use computer science to classify whales from sound recordings. This episode is about an evolution of that software by Atle Borsholm from NV5 Geospatial (https://www.nv5geospatialsoftware.com). At the beginning of the episode you will hear a sound recording. Take a listen and attempt to classify the whale making the calls in it. In my comments at the end of this I’ll tell you what whale that was. It was an easy one to process - the recording was just 7 seconds long and a whale sound was present immediately. Consider if I instead gave you another one lasting a day and most of it was just ocean noise. This leads to the challenge discussed in today’s episode. Imagine I have 100 microphones under the ocean. These are called hydrophones. These hydrophones record for 24 days. The scientists managing them will say they now have 2400 days of data to process. Well, that is nowhere near how big of a challenge this episode is about. The National Oceanic and Atmospheric Administration had 217,197 days of hydrophone recordings going back 20 years. Whilst Mark’s software was available, it needed to be made more efficient to process all this data. It was worth the effort. It turned out that only 24,000, or 11%, of those days involved a whale being detected. All of it is shown on NOAA’s Passive Acoustic Cetacean Map. Atle starts by discussing a web application, Mark’s Robots4Whales website, in our episode today. Links for further reading:NV5 article on Harnessing Data Science for Marine Conservation: https://www.nv5geospatialsoftware.com/Learn/Case-Studies/Case-Studies-Detail/harnessing-data-science-for-marine-conservationNV5 article on IDL® Software Extract Meaningful Visualizations From Complex Numerical Data: https://www.nv5geospatialsoftware.com/Products/IDLSlocum glider from Teledyne Marine: https://www.teledynemarine.com/brands/webb-research/slocum-gliderWoods Hole article on whale sound detection and classification: https://www.fisheries.noaa.gov/science-blog/listening-sounds-gulf The whale at the start? It was a humpback whale. Only Gemini managed to classify it correctly. Claude and ChatGPT got it wrong.
Transcribe →
Rebel Strategy Lab
2w ago00:29:22Tap to summarizeKarolina Sarna is founder of the wonderfully named Rebel Strategy Lab. She also writes and codes. On that basis alone I would want to record an episode. Even better though, she is an alumnus of the European Space Agency and Iceye, meaning she has a huge amount to contribute on the subject of earth observation. Iceye, by the way, have already come on the show.So, we covered:- How her background created an appetite for demand side thinking.- The defence pivot in commercial EO — why it's happening, what it's costing the climate/humanitarian side, and whether there's a structural way back.- Why "strong technology" keeps getting cited as the reason companies don't scale, and why that's usually wrong.- Her views on the Geoawesome Top 100.There was some issues with the internet, please forgive us for the occasional bad sound.So, my main reflection here is that it is possible to pivot out of industry into strategy consulting. Karolina had a great sense of humor and used this to deliver sensible points about the need for a constant focus on current problems in the industry you want to serve. Your product needs to help customers with those problems to make progress. She accompanied this message with another constant - stop focusing on the amazing technology involved in launching a satellite. This does not automatically mean you are solving someone’s problem. Earth observation satellites in fact deliver mostly noise. She reinforced this idea to the end, with a straightforward message about the limits of industry awards. Instead of the Geoawesome Top 100, we should instead focus on other industries identifying something great about a geospatial product. Then wait for that industry outside geospatial give an award to the company that makes this useful product.
Transcribe →
GIS at a Nigerian Power Utility and Earth Observation at Proforce Galaxies
3w ago00:57:00Tap to summarizeAbdulfatai Sanusi is a geospatial developer at Proforce Galaxies. He tells us the story of an MS Azure to Esri Enterprise migration at a power utility that covers several states in Nigeria. He then moves on to the next phase of his career at Proforce. They are an earth observation satellite design and data pipeline company. He described being a member of a talented team that also develops software. For example they have their own mobile data collection app and we discussed how this could actually be used in a utility context as part of the System of Monitoring. Overall it was a very eye opening discussion and showed us that the geospatial industry is alive an well in Nigeria! I was inspired, and went fishing in the Nigerian Stock Exchange for geospatial stocks to add to the GEO500. I found 3: Julius BergerCaverton OffshoreMTN NigeriaSo it was a fruitful exercise on all fronts, thank you Abdulfatai and please remember me for the Proforce IPO!
Transcribe →
QuantAgri - predictive agricultural insights
Apr 2800:36:56Tap to summarizeRyan Kmetz is back to talk with us about https://quantagri.substack.com/. This initiative began in January 2026. It is about analyzing satellite data to predict the monthly World Agricultural Supply and Demand Estimates (WASDE) reports. Ryan proposes a novel approach and shares some eye raising stats about lead times and profitability. This episode is inspiration for how we are all equipped to process the world using free data and AI. What can you do with your geospatial basic training?Links to items discussed:SatYield https://www.satyield.com/Sentinel ESA https://sentinels.copernicus.eu/web/sentinel/copernicus/sentinel-2Google Earth Engine https://developers.google.com/earth-engine/datasets/catalog/COPERNICUS_S2_SR_HARMONIZED#descriptionLandsat NASA https://science.nasa.gov/mission/landsat/Google Earth Engine https://developers.google.com/earth-engine/datasets/catalog/landsatTaylor Geospatial Engine field boundariesBackground https://taylorgeospatial.org/agricultural-field-boundaries-mapped-globally-for-the-first-time/Web app https://fieldsofthe.world/Field boundaries from Planet https://docs.planet.com/data/planetary-variables/field-boundaries/‘As the world turns’ Blackrock geospatial podcast https://www.blackrock.com/us/individual/podcasts/the-bid/geospatial-dataNovi https://www.geospatial.fm/p/novi-and-a-new-venture-geospatial
Transcribe →
Openclaw for Geospatial Monitoring Dashboards and Behavioral Investing
Mar 1001:05:55Tap to summarizeThis was quite the episode! We return for Round 2 with Nelson Roque. He is Assistant Professor of Human Development and Family Studies at Penn State. He walked us through OpenClaw. Nelson is an incredibly useful guest because he is able to cover both podcasts I have run - The Behavioral Investor and this one, Geospatial FM. A significant message from The Behavioral Investor was the effect of delay discounting or hyperbolic discounting. This is the decay in appeal of even fantastic outcomes when there is a significant delay until they happen. A formula is available here and here. Applying the formula shows that receiving a billion dollars in 3 generations, 108 years from now, only feels like $10,000 right now. The feeling decays with delay.There are a couple of ways to defeat hyperbolic discounting. One is episodic future thinking. This is a way to juice the dopamine reward system by a multimodal, multisensory imagining of what one might see, hear, smell, think, touch and feel upon achieving a financial goal. Turns out that, of course, AI has been applied to help here. Another that I came up with is instant, constant feedback about the reduction in amount of money one can pass to the next generation with each spending decision you make. What do I mean?Let's work with the $600 you might splurge on a smartwatch. The US market has compounded at 7% annually, adjusted for inflation and including dividends, the past 150 years. Imagine you are 42 and have a life expectancy of 72. You've therefore got 30 years of compounding $600 at 7% annually. This becomes $4,500 if you invest it in a low cost US market index fund with a 0.05% fee. So, you could enjoy the $600 watch or consider that in splurging on it you are effectively saying to your child you don't want to give them $4,500 when you die. So what I'm proposing is an app to make the effect of long time periods on investing outcomes visible. Please watch the episode as Nelson walks through producing this live with an AGI agent orchestrator called OpenClaw. He also profiles some geospatial situation monitoring apps (e.g. here). These were inspired by the incredible Bilawal Sidhu. He was able to give a geospatial data replay of the Iran strikes: here and here. It is amazing how far we've come since Nelson's sceptical comments in the first episode with Nelson 9 months ago about whether or not we have AGI. Let's check back on his opinion of AGI in another 9 months.
Transcribe →
Geoharbor - Hearding the Data Engineering Cats
Mar 100:52:11Tap to summarizeGeoharbor is a solution from SketchMyView. It is a way to avoid notebook hell. This is done through all data engineers on a team using a common framework. I recommend a watch to see Rakesh showing it in action.
Transcribe →
Seerab Technologies on How to Win Uber Tech Prizes
Feb 1000:34:33Tap to summarizeWe are privileged today to share an episode with a Pakistani geospatial firm, Seerab Technologies. Zulkifil is a co-founder and has been with the company nearly 10 years. They have their own platform for real estate developers to help customers understand the context of proposed property purchases. He gave a simple example, a property with a large hole in it is hard to spot if you don’t have a terrain map. Seerab provides that to users. He also stepped through a few prizes they have won, including from Uber. Nice to have some lessons learnt from a firm all the way over in Pakistan telling us how product development is done. What was most interesting to hear is the timeless advice he shared applies everywhere. We’re so used to hearing it from Silicon Valley, but his is a story of applying it successfully a world away - Seerab has been in business for years. Finally, Zulkifil talks about doing good. According to Gemini:"The Hajj is one of the world's largest annual gatherings, with over 1.8 million to 2.5 million pilgrims converging on Mecca, Saudi Arabia, for five to six days of rituals. It is a mandatory, once-in-a-lifetime duty for all adult Muslims who are physically and financially capable."One may imagine with so many people scrambling across a relatively small area on foot, in an unfamiliar place, there is a risk of getting lost. So Zulkifil showed us a solution based on listening to Pakistani pilgrims going to the Hajj.Thanks Zulkifil for taking the time!
Transcribe →
Thoughts on Weather Foundation Models with Alex Merose 3/3
Jan 3000:39:54Tap to summarizeThis is episode 3/3 with Alex Merose about his thoughts on weather foundation models. He is a member of the technical staff at Open Athena. In this episode, Alex steps through system characteristics of weather foundation models and how we can approach building them. Toward the end, the episode touches on an example of applying these approaches to a simulation of the Earth’s weather over a period of 800 years. The approaches Alex has been talking about enable the use of a GPU to process this simulation in only one day. We conclude with the values that drive Alex’s work.
Transcribe →
Thoughts on Weather Foundation Models with Alex Merose 2/3
Jan 2100:46:25Tap to summarizeThis is episode 2/3 with Alex Merose about his thoughts on weather foundation models. He is a member of the technical staff at Open Athena In this episode, Alex steps through ML weather models, their history and how they work.Links to items discussed:ECMWF Reanalysis v5 (ERA5): https://www.ecmwf.int/en/forecasts/dataset/ecmwf-reanalysis-v5I got fooled by AI-for-science hype—here's what it taught me: https://www.understandingai.org/p/i-got-fooled-by-ai-for-science-hypeheres
Transcribe →
Thoughts on Weather Foundation Models with Alex Merose 1/3
Jan 1300:29:20Tap to summarizeThis is episode 1/3 with Alex Merose about his thoughts on weather foundation models. He is a member of the technical staff at Open Athena. In this episode, Alex steps through the background on doing weather prediction, from early efforts around a century ago to numerical and physics based models. This prepares us for later episodes on machine learning or AI based approaches. Links to items discussed:Pangeo, a community for open, reproducible, scalable geoscience.Alex's Google Scholar profile.Episode with Sergei Nozdrenkov on a coral reef foundation model.Global Climate Data Collaboration: The Intentional Dream by George Dyson.Hurricane Melissa.Arham Ansari on GeoRiskAI.What can a technologist do about climate change.
Transcribe →