Wherobots

What is Wherobots? It's a question we get asked sometimes, so we wanted to make it super easy to understand. That's why we created this handy explainer video. Check it out here 👇

🚗💡 What if your geospatial models could think in travel time, not just distance? Join us for a live session where we're introducing a new Isochrones Dataset—a powerful, open-source layer that maps reachable areas based on time and transport mode, built on the Overture Maps POI Dataset. 🗓️ Webinar: How Wherobots Generated Drive Time Isochrones for Every POI in the US Topics we'll cover: 🔹 What makes the Isochrones Dataset unique and how it was built 🔹 How to integrate it with demographic, land use, or POI datasets 🔹 A hands-on demo of geospatial weighted regression to model service coverage & accessibility 🔹 Why time-based modeling beats distance-based approaches in urban analytics Whether you're in urban planning, retail strategy, healthcare accessibility, transport modeling, or any spatial domain where time matters—you’ll gain insights and tools to level up your spatial analysis. 👉 Save your spot: https://lnkd.in/dunA8q92

Happy Earth Day! 🌎 We're excited to share that WherobotsAI Raster Inference now supports Meta’s Segment Anything 2 (SAM2) model! You can now perform object detection and feature segmentation at scale on terabytes of satellite imagery—just by describing what you’re looking for in plain text. 🔤 Ask questions like: “Find the container ships” 🚢 or “Segment pickleball courts” 🎾 across very large areas of interest. Because these results are stored as Iceberg tables in your S3 bucket, you can easily join with other datasets using WherobotsDB and continue your analysis inline with 300+ spatial features and functions. 🔍 Read more about text to detection in satellite imagery: https://bit.ly/4irLlsw

Just a friendly reminder that our live webinar on zonal stats is tomorrow! Don't forget to register, and we’ll see you there! 👇

Heading to the Cloud-Native Geospatial Forum (CNG) Conference next week? Learn how to leverage ML to extract deeper insights from satellite imagery — you won’t want to miss this session led by Damian Wylie. Day 2 | 11:45am-1:15pm 📍 Wasatch A | Getting Started with Cloud-Native Geospatial Workflows 1: Solutions 🔍 Session: Extract insights from satellite imagery at scale with WherobotsAI Inferring objects and detecting change in satellite imagery was once reserved for companies with the talent, money, and time to build, manage, and run sophisticated, self-managed ML inference solutions against satellite data.

Dotlas tackled a spatial join task involving restaurant data from multiple sources. After conflation, the dataset contained approximately 1.1 million restaurants. 🍽️📍 With BigQuery, the data had to be moved to GCP and partitioned by state, resulting in a runtime of about 4.5 hours. Using Wherobots, however, the data remains entirely within AWS and partition it any way you want—reducing the runtime to just 18 minutes. ⚡ Hear it from Dotlas Co-Founder and CTO Eshwaran Venkat ⎋ himself. 👇

🚒 Imagine you're tasked with identifying the next location for a new fire department across the entire U.S., based on clusters of high-risk areas. How would you approach that? 💭 💡You can do this by using isochrones in combination with 13.3 million Overture Places. 🚀 For a total compute cost of less than $110 and in under 30 minutes of processing time, you can answer and automate nationwide questions on-demand, such as “Where is the risk?” or “Where should my [investment or asset] be located in the country?” And no, this isn’t limited to fire departments—this approach is valuable for anyone working with location intelligence. ➡️ Follow along to see how we did it: https://bit.ly/4lnE93m

Zonal statistics is one of the most powerful ways to combine raster and vector data—but once you try to scale beyond a few polygons, most tools start to struggle. Join us for a live session where we’ll show how Wherobots make large-scale zonal stats simple, fast, and scalable for modern spatial SQL. 🔍 What we’ll cover: ✅ How to prep and load raster + vector data ✅ Using RS_ZonalStats to compute mean, min, max & more ✅ Real-world demo using Overture Maps + AWS elevation data ✅ Tips for optimizing and scaling your raster analysis workflows Perfect for anyone working with terrain, land cover, or climate data—this is your shortcut to efficient, scalable raster analysis. 👉 Reserve your spot: https://bit.ly/4jzp5xQ

💡 Did you know that using GeoPandas along with Wherobots and Apache Sedona can provide powerful, efficient analysis, especially for large datasets? Learn how to: ✅ Run computations on large datasets in parallel with Wherobots ✅ Use GeoPandas for detailed analysis and visualization ✅ Write Sedona code using familiar GeoPandas syntax ✅ Seamlessly switch between engines depending on your needs Whether you're scaling to massive datasets or diving into detailed analysis, this blog shows how these tools ✨work better together✨. https://bit.ly/42devXG

⏰ Last week I shared a quick benchmark: Wherobots ran a population enrichment query in 57 seconds, BigQuery took 1 minute 20 seconds... But I wanted to dig into the why, because the difference isn’t just about speed. It’s about architecture. This was a complex area-weighted interpolation - geometry-heavy and compute-intensive. And that’s where purpose-built spatial systems shine. In this new doc, I break down: ✅ How Wherobots handles geometry differently (spatial indexes + parallelism) ✅ Why working directly with GeoParquet and Iceberg saves a lot of pain ✅ What true data lakehouse architecture looks like for spatial If you’re working with large geospatial data and tired of complex pipelines and ETL gymnastics this is for you. 👉 Swipe through the slides for a deeper look. #gis #moderngis #geospatial #cloudnative #apacheiceberg