Geospatial FM

Michael Proulx from Eidolon Strategic ESG is the next in line contributing to the mine tailings dam issue. Michael makes meaningful progress for us on tailings dams because he used to lead a bond trading desk and is now an ESG analyst. Hence, we can connect three threads: how money markets work, how a disaster reaches a stock price, and where geospatial data fits in. We start by walking through the basics of equity versus credit markets. Equity holders bet on growth, while bondholders prioritize getting their principal back. This is why bondholders sit higher in the capital stack and recover more when a company fails.The conversation then follows a concrete case: the 2019 Brumadinho tailings dam collapse in Brazil, and the 28% drop in Vale’s share price that followed. This is measured from the closing price of the prior week, a Thursday, because the event happened on a Monday. The comparison period ends February 8th 2019, as shown above. Over those 10 trading days, the value of the company went from $USD77B to 58B. A nigh on 30% drop in the value of a company, especially when it amounts to $20B lost, warrants attention. Michael explains how a rating agency’s event-scoring system can trigger institutional investors’ committees to sell, and how the resulting selling pushes the price down. Event-driven selling, not only the disaster itself, is what moves the market.The discussion then turns to what this means for mining oversight. Pension funds in the UK and Sweden pushed listed miners to disclose tailings facility locations, creating a dataset of roughly 2,000 sites. The discussion covers how geospatial analysis could go further: measuring proximity to communities, density changes, and early signs of instability.The episode closes on conflict risk. Michael’s master’s research connected mining to conflict, and he points to cases where armed groups take over mines, drop maintenance, and leave tailings unmonitored. He also suggests that satellite imagery could help track armed group movements, and that geospatial intelligence tools might give investors earlier warning.Listen to the full episode for the worked examples and the discussion of how to turn these risks into decisions.

My guest is Silas Toms, GIS supervisor at PG&E, which covers the largest land area of any utility in California. Silas has 20 years in geospatial, including time at Zillow, and runs GeoMeetup in SF. Yes, he was the one wearing that tracksuit. The next one is tomorrow.We cover why utilities should separate their systems of record, engagement and monitoring, and how AI changes who builds geospatial tools. We also look at GIS Map Online, Silas's open-source, browser-based replacement for Google Earth.Stick around to the end for a 10 minute demo.

Kevin T. Carter is founder of a couple of investment firms, EMQQ Global and EMXETF. You’ve probably heard of the Magnificent 7. These are stocks from what the West calls Developed Markets. Kevin talked about applying the same idea to emerging markets, he found The MAG 3.Then we moved on to China’s AI Tigers. Kevin challenged our assumptions about US AI leadership with stories from China. Finally, we covered how to create an exchange traded fund or ETF. This matters because there isn’t one for geospatial yet.

Today’s episode is with the founder of Dekart, Volodymyr Bilonenko.We cover WHY he made Dekart, the relationship with Kepler, who he competes with and some discussion of specific geospatial needs of mobility companies like Wayve.He then steps through use of a new Claude skill he made called geoSQL.This skill uses a map in the loop, yes Claude checks images of its own maps in the course of producing a final Dekart output.We finish the discussion by looking at how geospatial business models may evolve in the context of Claude native GIS.

Ankur Shah, https://www.linkedin.com/in/ankur-s-803497126/, is founder and CEO of PlanetSapling (https://planetsapling.com/). He talks us through a dataset of 3,000+ mines he's found the location of, and ownership information about, worldwide. He also takes us through related analyses he's been doing, such as with nature data company Cecil (https://www.linkedin.com/pulse/mineral-supply-chain-nature-risk-screening-cecil-planetsapling-7rqmc/). Finally, we turn our attention to the problem of mismanaged mine tailings storage facilities. Or tailings dams. Every year there is a notable collapse of one. An example he gives is from last year where a Sino-Metals Leach Zambia tailings dam collapsed in Zambia. Toxic waste hit the Mwambashi River. This flows into the Kafue River and 60% of Zambians live in its river basin. Roughly 45% of Zambia's population relies directly on the Kafue River for their drinking water. That is 10 million people. Join us today to contribute to the development of a global, authoritative, continuously updated dataset of tailings dam boundaries. Your expertise is sought in creating it, then standing up a tailings dam runout modelling engine. Its outputs regarding predicted lives lost and critical infrastructure damage will incentivise appropriate action by leaders. Or discipline of those leaders by the population.

Weekly wrap for 7 days to 7/26/2026 and promo for Quant Agri (https://quantagri.substack.com/) with Ryan Kmetz (https://www.linkedin.com/in/ryankmetz/). I will be incorporating Quant Agri GEOINT-based commodity trading performance going forward, based on a $1,000 starting position.

marimo is the fastest growing Python project with 22,000 Github stars. Fantastic episode today, covering:0:00 - Intro0:54 - Favorite keyboard2:35 - What Coreweave's GPUs mean for marimo, such as solving the Travelling Salesman Problem8:13 - What is the one thing Jupyter Notebooks always gets wrong, but marimo doesn't?17:43 - UI elements in marimo library like sliders, cascading updates across cells and marimo notebooks as web apps20:08 - What marimo means for different modern data team personas 24:39 - Remote storage connections and example of data analyst work27:15 - How marimo is used by various scientific disciplines and using Anywidget in marimo31:45 - Use of marimo in an enterprise context for GPU compute, roadmap comments and creative freedom the tool enables35:37 - Using marimo to defeat the Esri monopoly in geospatial: if you are David, play a different game than Goliath, again the point that marimo/molab offers creative freedom to sponaneously poke around in the data42:27 - How to use GPUs in marimo, limitations of this47:55 - Using AI to control marimo 50:30 - Call to action - marimo competition to make coolest notebook from an academic paper51:20 - Vincent's summary: this is marimo, go and be David, fight Goliath, most of all - creative freedom. [We are trying] to figure out a way for you to be more creatively free.51:35 - Dice kit55:48 - More on keyboards.57:25 - I try pairing Claude and marimo myself.A huge privilege to have Vincent on to tell us all about it. Here are some links to items discussed:marimo joining Coreweave: https://marimo.io/blog/joining-coreweavemarimo repo: https://github.com/marimo-team/marimomarimo pair - for helping pair with an AGI: https://github.com/marimo-team/marimo-pairA related blog post: https://marimo.io/blog/claude-codeEvidence marimo uses Leafmap: https://docs.marimo.io/api/plotting/?h=map#other-plotting-librariesAnywidget: https://anywidget.dev/ and its dev: https://github.com/manztNotebooks in Wherobots: https://wherobots.com/blog/wherobots-notebook-environment-getting-started-with-wherobots-cloud-sedonadb-part-2/Vincent's Wiggly Stuff: https://koaning.github.io/wigglystuff/marimo recently ended competition: https://marimo.io/pages/events/notebook-competition-2We are now discussing the next one and adding a geospatial theme to it!

Guest: John Metzger (LinkedIn), Founder of Asset Assurance Monitoring (Website)."I don't know how, but I recognized it was one of my children. My friend rescued Nicolas, and I thought: Oh my God, at least I managed to catch one."These were the words of a father fleeing the 2015 Bento Rodrigues (Mariana) dam disaster in Brazil (O Globo, 08/11/2015). Five days later, his daughter Emanuele was found dead. The collapse killed 19 people and sent a torrent of toxic mine waste 670km down the Doce River to the Atlantic Ocean. It was Brazil’s worst environmental disaster.Four years later, just 70km west, it happened again. The 2019 Brumadinho dam collapse killed 272 people—including an entire family of five and an unborn child—becoming Brazil’s worst industrial disaster. Both mines were owned by corporate giant Vale.When raw ore is extracted, the toxic, liquid byproduct is stored behind massive earthen structures called tailings dams. Globally, there are an estimated 29,000 to 35,000 active, inactive, or abandoned tailings storage facilities (TSFs) holding 223 billion tonnes of waste (World Mine Tailings Failures). Active sites account for 85% of all failures. The risk is ongoing: as of late 2025, Brazil alone has 916 dams, with 74 at high risk of collapse and 91 on alert, concentrated heavily in the mining hub of Minas Gerais.When these structures are poorly designed or neglected, they fail. When they fail, the wall of mud obliterates everything in its path.Our guest today, John Metzger, is an expert on the advanced monitoring systems designed to prevent these catastrophes. Having led an incredible, multi-country career, John joins us to explain how real-time data, geotechnical instrumentation, and rigorous telemetry save lives.There were fatal systemic flaws of recent disasters—including why Brumadinho’s emergency warning alarms failed to sound, echoing previous conversations on the show with Floodmapp’s Juliette Murphy on the desperate need for strict flood-alarm regulations (YouTube Link). Also corporate failures: reports that Vale knew of automated sensor malfunctions two days before the Brumadinho collapse (Mining.com Report) and allegations that safety inspectors felt corporate pressure to sign off on unstable structures.Despite these failures, there are rays of hope. First, the recent establishment of the UN-backed Global Tailings Management Institute (GTMI), a new international watchdog tasked with ending corporate negligence in tailings management. Second, I want to honor the legacy of Lindsay Newland Bowker, the actuary behind the vital World Mine Tailings Failures database, who passed away in May 2026. To ensure her work is not lost to time, I have an upcoming conversation with Ankur Shah of PlanetSapling to discuss the future of open-source risk mapping.

I am joined by Robert Cheatham of Japan Earth Observer (https://www.japanearthobserver.com/) for a GEO500 Weekly Wrap. Provisionally, also the PhysicalAI Weekly Wrap. I'll think about it. GEO500 weekly roundup sheet: https://docs.google.com/spreadsheets/d/1AVbN5pQ9kiRmC0Iw2H8stoPJUngrqntA5DRU3Wll2aE/edit?usp=sharingTacticAI: Google's AI system that can help simulate field scenarios and predict open play dynamics up to 8 seconds in advance. ⚽ https://www.linkedin.com/feed/update/urn:li:share:7470859786159370240/Google Earth infrastructure layers: https://x.com/googleearth/status/2064752086386930152Oracle's disastrous quarter: https://www.tradingview.com/news/zacks:a8c6c958c094b:0-oracle-shares-tank-despite-q4-earnings-beat-hold-the-stock-now/CAD/BIM vertical destruction:Claude Fable 5 has solved CAD: https://x.com/aaronli/status/2064876123109089742?s=46Nvidia is coming for Autodesk: https://x.com/nvidiartxspark/status/2061825397868630454?s=46Claud Fable 5 makes 3D/FPS video games in minutes: https://x.com/chrissgpt/status/2065193150222663959?s=46SpaceX jobs: https://www.spacex.com/careers/jobs

Today's guest is Robert Cheatham https://www.linkedin.com/in/rcheetham/ and we are talking about Japan Earth Observer, https://www.japanearthobserver.com/. This is a newsletter on the space, Earth observation, and geospatial industry in Japan. It is based on a fantastic effort by Robert to systematically collate every last company in the space and earth observation supply chain in that country. Naturally I was drawn to it due to my efforts to do this not only for Japan but every country in the world as the GEO500. We got in touch and exchanged our lists which led to the discussion recorded here today.The GEO500 has 56 Japanese companies since inception. 7 of them are delisted. So the index contains 49 live positions for this country. All GEO500 positions start with $100. The entire Japanese part of the GEO500 stands at $AUD13,000. This is because the 56 have compounded at 8% annually since inception. In the case of this country, starting as a position in Fujitsu in 2000. The top three performers in terms of capital growth rate are Terra Drone (125% annually), Synspective (71%) and QPS (59%). The top 3 in terms of value are Sony ($1100 - bear in mind the $100 starting position, equating to a 20% annual growth rate since entering the index in 2013), NEC ($960, 10%) and Raito Kogyo ($760, 11%). Note that the 3 fastest growers are all earth observation companies, and 2 of those are synthetic aperture radar satellite constellations. This is an exciting part of our industry to watch and I have recorded a couple of great episodes on the power of it. For example, - Jamon Van Den Hoek's work as an academic partnering with some of the world's largest media organisations to monitor building destruction in warzones and after fires: https://open.spotify.com/episode/0RQNjzDtTQTNqv8IiMy2SU- Umbra, a pioneering American SAR satellite manufacturer, they also operate their own constellation: https://open.spotify.com/episode/2QM3OeglXI1naiSgrKod17- Iceye, the equivalent from Finland: https://open.spotify.com/episode/04S3fsdL4cus2tOEbq4KBN- Ursa Space Systems, a commodity intelligence firm that uses SAR to monitor everything from iron ore stockpiles at ports to oil farm tank lids around the world: https://open.spotify.com/episode/3ltO0Bv8Mtq7dgnLhXWGxO- SeerAI, another analytical platform where I coaxed the guest to use SAR as a way to detect surface change on the NEOM excavation, the world's largest (now abandoned) building site: https://www.geospatial.fm/p/seerai-responds-to-johnny-harrisBut this wasn't just about current companies. We were privileged in this episode to be given a history of Japan's trading houses since the 1860s to today. This led to several insights around the longevity, for example, of the social groups that comprise such firms and how they can persist even through a trust busting effort by the winning country after a world war. Another insight was the benefit to society of trust busting and how it unleashed a wave of new companies and from that household names such as Honda and Sony. Honda recently launched a rocket that was able to land itself. This has relevance to possible future earth observation constellation launches. Sony released LiDAR solution in 2013 and also sells other geospatial products like a GNSS chip. We have come full circle.A final note about Robert himself. We are presented here with a cultivated and successful entrepreneur. He grew a geospatial software development firm to 50 people then sold it. He has learnt a foreign language well enough to spontaneously translate things for me during a podcast recording without warning. He also is able to offer a comprehensive historical view on the emergence, development, setbacks and modern day expression of several dominant companies in a vibrant foreign economy, Japan. He shows us the way, then, on multiple fronts. I am grateful for the chance to put an inspiring industry figure in front of you today. Thanks Robert.