Skip to content
Sign in
Artwork for Project Geospatial
Technology

Project Geospatial

Project Geospatial

Project Geospatial is a non-profit media organization committed to growth, innovation and collaboration among academic, commercial and government partners in the geospatial industry. We believe a revolution is coming in geospatial technology – big data is democratizing, satellites are shrinking, and the science of where has never been more important to the health and future of our planet. We provide a leading voice in geospatial industry trends, insights and cross-industry applications to help inspire and conspire with the next generation of great geospatial ideas.

Play
Open in
  • 55 episodes
  • Avg 12 min
  • English

Episodes 55

NewestOldestLongest
All20Unplayed20In progress0Played0
Counted on this page — what you have heard stays on this device, so it is not something the list can be paged by.
  • September 7 · 54 min

    FedGeoDay 2025 | Panel - Open Data for Resilience

    This panel, moderated by Eddie Pickle, brings together leaders in the open data community—Maggie Collie (OpenStreetMap US), John Crowley (MAPAP Action), Angelina Calddron (Meta / Overture Maps Foundation), and Jed Sunwal (Radiant Earth)—to discuss the role of open data in fostering resilience. They highlight the importance of initiatives like OpenStreetMap (OSM) and Overture Maps in building communities, developing tools, and providing crucial data for emergency response and other applications. The discussion emphasizes that resilience relies not only on the availability of open data but critically on the human curators and institutions that maintain and update it, especially in rapidly changing environments. Panelists share success stories of collaborating with federal agencies, such as using OSM to fill gaps in federal transportation and trails data. They explore the challenges of sustaining open data efforts and engaging government effectively, touching on issues like licensing, bureaucracy, and ensuring feedback loops from data users back to data creators. The conversation stresses the necessity of durable institutions, aligning incentives across diverse stakeholders, and fostering collaboration and community engagement as foundational elements for building a truly resilient geospatial data ecosystem. • Panel discusses the role of open data in promoting resilience. • Features leaders from OpenStreetMap US, MAPAP Action, Meta / Overture Maps Foundation, and Radiant Earth. • Highlights key efforts like OpenStreetMap (OSM), Overture Maps, and Humanitarian OpenStreetMap Team (HOT). • Emphasizes that resilience depends not just on open data availability but critically on the human curators and institutions that maintain and update the data. • Discusses the importance of building durable institutions and aligning incentives across diverse stakeholders (users, creators, companies, government) to sustain open data efforts. • OpenStreetMap US focuses on supporting the community, tools, and education to lower barriers to contributing and using data. • MAPAP Action uses open data extensively for international disaster response, stressing the need for updated data in dynamic situations. • Meta / Overture Maps contributes and uses open data, aiming for data quality, usability, and scale, particularly with the Global Entity Reference System (GURS). • Radiant Earth works to make free data "cheaper" by improving web-accessibility and ergonomics of large datasets using cloud-native formats. • Shares success stories of federal agency collaboration, such as using OSM to fill data gaps in transportation (BTS) and trails (NPS). • Touches on challenges in government engagement with open source/data due to issues like licensing (ODbL), bureaucracy, trust, and funding models. • Advocates for feedback loops from data users back to data creators/maintainers to improve data quality over time. • Highlights the fragmented nature of the US domestic disaster response architecture (federal/state/local GIS capacity) and the need for rethinking. • Emphasizes the need for collaboration and community engagement across government, industry, academia, and non-profits as fundamental for building a resilient ecosystem. • Suggests focusing on community-led standards/specifications and aligning around common formats and identifiers (like GeoParquet, STAC, GURS). • Concludes by stressing the importance of partnerships, staying connected, and community as crucial for the future of open data for resilience. For more content like this check out www.projectgeospatial.com #FedGeoDay #Geospatial #OpenData #Resilience #OpenStreetMap #OSM #OvertureMaps #MAPAPAction #RadiantEarth #Community #Collaboration #HumanitarianResponse #DisasterResponse #DataCuration #FederalGovernment #GIS #CloudNative

  • September 7 · 53 min

    FedGeoDay 2025 | Panel - Demystifying AI

    This panel, moderated by Bill Doins, features experts Emily Cala, Jason Gilman, Ron Goldblat, and Jackie Kaisle discussing the role of AI in the federal geospatial landscape and demystifying common terminology. Panelists share diverse applications of AI they are working on, from detecting insights and optimizing workflows to image analysis for damage assessment and predicting disease outbreaks. They address the often vague definitions of terms like AI, machine learning, and LLMs, emphasizing the need for clarity and a "problem-first" approach. Challenges and opportunities with federal geospatial data include acquiring sufficient labeled data (discussing synthetic data and newer models), ensuring data quality, and using open standards. Ethical considerations are paramount, focusing on privacy, addressing biases (geographical, temporal, social/demographic), and the need for transparency and explainability in AI outputs. The panel stresses that AI is a tool requiring careful implementation and validation, not a magical shortcut, particularly in critical government applications where inaccuracies can have serious consequences. They discuss the relatively slow pace of federal AI adoption but see opportunities for time savings and leveraging collaboration across government, academia, and industry, recommending focus on proven solutions and community-led specifications rather than chasing every new trend. Ensuring accuracy for critical applications may involve methods like model factories and human-in-the-loop verification, while security concerns like "rag poisoning" highlight the need for trusted data sources and rigorous testing. • Panel discusses the use of AI in the federal geospatial space and aims to demystify related terminology. • Panelists share diverse applications including insights detection, workflow optimization, image analysis for damage assessment, and disease prediction. • Discusses the often vague definitions of AI, machine learning, LLMs, etc., and emphasizes the need for clarity. • Advocates for a "problem-first, model-second" approach to AI adoption. • Highlights challenges with federal geospatial data for AI, such as the need for labeled data, metadata quality, and leveraging open standards. • Explores concepts like synthetic data and newer zero-shot/one-shot models to address labeled data scarcity. • Emphasizes critical ethical considerations: privacy concerns (protecting personally identifiable information), addressing biases (geographical, temporal, social/demographic), and the need for transparency and explanability in AI outputs. • Stresses that AI is a tool requiring thoughtful understanding and implementation, not a magic shortcut. • Notes the relatively slow pace of AI adoption in the federal government but sees opportunities for time savings and efficiencies. • Discusses ensuring accuracy for critical applications, requiring validation, potentially using model factories and human verification. • Addresses security concerns, including the concept of "rag poisoning" for LLMs, highlighting the importance of trusted data sources and rigorous testing. • Suggests cutting through AI hype by focusing on proven, practical solutions. • Recommends collaboration across government, academia, and industry as crucial for advancing AI adoption. • Discusses the tension between setting standards and the rapid pace of AI development, suggesting community-led specifications (like STAC) might be more effective than rigid standards. For more content like this check out www.projectgeospatial.com #FedGeoDay #Geospatial #AI #MachineLearning #LLMs #ArtificialIntelligence #FederalGovernment #Resilience #EmergencyManagement #OpenSource #OpenData #Ethics #Bias #Cybersecurity #DataValidation #Collaboration #GIS

  • September 7 · 6 min

    FedGeoDay 2025 | Lightning Talk - Shane Brennan

    Shane Brennan discusses a method for classifying 3D point clouds by leveraging 2D computer vision techniques. He focuses on point cloud data collected from moving vehicles equipped with LiDAR sensors, explaining that this data can be represented as panoramic images containing information like range. The core idea is to apply existing 2D object detection and segmentation models, such as YOLO, to these panoramic images to identify and classify objects at the pixel level. Because the pixel location in the panoramic image can be accurately mapped back to the original 3D point cloud using the range and angular measurements, the 2D classification can be efficiently propagated to the corresponding 3D points. Shane highlights that this approach is particularly effective when integrated with Simultaneous Localization and Mapping (SLAM) processes, allowing for classification while mapping the environment. He notes that this method offers an advantage over traditional 3D point cloud deep learning techniques, which often require large amounts of difficult-to-produce labeled 3D training data. • Discusses a method for classifying 3D point clouds using 2D computer vision techniques. • Focuses on point cloud data acquired from moving vehicles equipped with LiDAR sensors. • Explains that LiDAR data can be represented as panoramic images containing useful information like range. • Proposes applying 2D object detection and segmentation models (like YOLO) to these panoramic images to classify pixels. • Highlights the ability to propagate 2D pixel classifications back to the original 3D point cloud using range and angular measurements. • Notes that this method is particularly useful in the context of Simultaneous Localization and Mapping (SLAM). • Presents this approach as an advantage over 3D deep learning methods that require difficult-to-produce labeled 3D data. • Mentions the use of a structured LiDAR sensor like the Ouster OS1. For more content like this check out www.projectgeospatial.com #FedGeoDay #Geospatial #PointClouds #LiDAR #ComputerVision #MachineLearning #AI #YOLO #SLAM #3DData #ObjectDetection #Segmentation

  • September 7 · 5 min

    FedGeoDay 2025 | Lightning Talk - Ryan Burley

    Ryan Burley from GeoSolutions discusses their involvement in developing and supporting open source geospatial software and applications for climate resilience and emergency management. GeoSolutions is a key contributor and steering committee member for prominent open source projects like GeoServer, GeoNode, and MapStore. They support various customers, including those in critical domains such as Sea Island Software (Hurrvac), NOAA (NowCoast for near-realtime OGC services via GeoServer), and Critical Response Group (CRG), providing customized GeoServer solutions to deliver floor plans and raster data quickly to first responders. GeoSolutions specializes in managing and delivering large, complex spatiotemporal data from sensors to end-user applications, working across diverse sectors. He also highlights MapStore as a flexible web mapping framework for visualization and dashboards and mentions their expertise in Cesium 3D Tiles. Ryan provides an update on the upcoming GeoServer 3, a significant crowdfunded upgrade to modernize its underlying technologies, expected to be released in late 2025. • Discusses GeoSolutions' role in developing and supporting open source geospatial software for climate resilience and emergency management. • GeoSolutions provides core code developers and steering committee members for projects like GeoServer, GeoNode, and MapStore. • Supports various customers, including Sea Island Software (Hurrvac), NOAA's NowCoast service (using GeoServer for OGC services), and Critical Response Group (CRG) (custom GeoServer for first responders). • Specializes in the ingest and delivery of complex, large spatiotemporal data from sensor to end-user applications. • Works across diverse industries, including maritime security, smart cities, precision agriculture, transportation, and energy. • Highlights MapStore as a versatile web mapping framework for front-end visualization and dashboarding. • Mentions GeoSolutions is a Cesium certified developer organization with expertise in 3D Tiles. • Provides an update on the major upgrade to GeoServer 3, a crowdfunded effort expected for release in late 2025, which modernizes its security framework and JDK version. • Lists various online resources for learning more about GeoSolutions and these open source projects. For more content like this check out www.projectgeospatial.com #FedGeoDay #Geospatial #OpenSource #Resilience #EmergencyManagement #GeoServer #GeoNode #MapStore #GeoSolutions #OGC #SpatialData #ClimateResilience #CriticalResponseGroup #NOAA #Hurrvac

  • September 7 · 5 min

    FedGeoDay 2025 | Lightning Talk - Quincy Morgan

    Quincy Morgan introduces Layer Cake, a new cloud-native tool from OpenStreetMap US providing access to OpenStreetMap (OSM) data in the GeoParquet format. He highlights OpenStreetMap as the world's largest free, editable vector data repository, fostering collaboration and digital democracy. OpenStreetMap US supports this ecosystem in the United States. Layer Cake offers a direct and easy way to access this data, currently providing weekly updates for building and highway layers. A key feature is exposing significantly more raw OSM tag data than many other tools and containing only OSM data, offering users a closer connection to the source. The GeoParquet format facilitates easy querying and integration with various tools like DuckDB, QGIS, and ArcGIS Analytics. The objective behind Layer Cake is to increase OSM adoption and usage, thereby supporting digital democracy and building resilience by giving users control over non-proprietary data. OpenStreetMap US also supports various other tools and initiatives, including a planned vector tile service. • Introduces Layer Cake, a new cloud-native tool from OpenStreetMap US for accessing OSM data. • Provides access to OpenStreetMap (OSM) data in GeoParquet format. • Highlights OSM as the world's largest repository of free, editable vector data, promoting collaboration and digital democracy. • Offers a direct and easy way to query OSM data, currently providing buildings and highways layers with weekly updates. • Exposes significantly more raw OSM tag data compared to other solutions. • Contains only OSM data, with no third-party additions. • The GeoParquet format enables easy integration and querying with tools like DuckDB, QGIS, Pandas, and ArcGIS Analytics. • Aims to support OSM adoption and usage, which helps build resiliency by providing control over non-proprietary data. • Mentions OpenStreetMap US also supports tools like MapRoulette, OSMcha, and Field Papers, and has working groups for collaboration. • Announces a coming vector tile service to make using OSM easier in web maps. For more content like this check out www.projectgeospatial.com #FedGeoDay #Geospatial #OpenSource #OpenStreetMap #OSM #LayerCake #GeoParquet #CloudNative #VectorData #DataAccess #Resilience #DigitalDemocracy #Community

Showing 1–20 of 55 episodes