Skip to main content

​
Start With Your Question

Spatial work usually starts with a business question. Open your industry to see the questions Wherobots answers, what each one produces, and the notebook that implements it.
“Where has the land changed since our last survey, and how much?”You get vectorized model outputs for the interval between two dates, by running a change detection model over satellite imagery from both dates. The example notebook shows the two-date workflow with the Fields of the World model, which segments field boundaries rather than detecting change. To measure change, run a change detection model, such as one you export with the Bring Your Own Model notebook.

Change detection with RasterFlow

Run a model over imagery from two dates.
“Which of our operations fall inside protected areas or hazard zones?”You get a flag on every operational boundary that overlaps a protected area, waterway, or hazard zone, from a single spatial join against regulatory boundaries.

Spatial joins in Wherobots

Join asset footprints against boundary datasets at scale.
“Where are the solar panels, parking areas, and other roof and site features across our portfolio?”You get per-building detections with confidence scores, joined back to your asset records, by describing the feature in a text prompt and running SAM3 geometry inference over sub-meter aerial imagery.

Detecting objects from text prompts

Describe a feature in words and get georeferenced geometries back.
“Which sites in our portfolio carry the most location risk?”You get the severe weather events, such as hail and storm cells, near each site, by joining site locations with NOAA storm records and other hazard layers such as flood zones.

Exploring storm data

Correlate NOAA severe weather records with your locations.
“How many customers can reach each potential new site in 15 minutes?”You get drive-time catchments for every candidate, with population and competitor counts inside each, by generating isochrones in batch and spatial-joining them to demographic and POI data.

Making drive time isochrones

Build travel-time polygons for many locations at once.
“Where are competitors densest, and where is the gap?”You get competitor counts and market saturation around every candidate site, by joining candidate locations against POI datasets and clustering the results.

Using Overture Maps

Work with POI data to map the competitive landscape.

DBSCAN clustering

Find natural groupings in demand and competitor locations.
“Which roads did each vehicle actually drive?”You get every trip snapped to real road segments, by map-matching raw GPS traces against the road network at fleet scale. Compare the matched paths with planned routes to find deviations.

Matching vehicle GPS data to road networks

Snap noisy GPS traces to the roads they were actually driven on.
“Which trips look anomalous and deserve a second look?”You get a ranked list of outlier trips and GPS points, by applying Local Outlier Factor to your trajectory data.

Local Outlier Factor

Surface spatial outliers that summary statistics hide.
“What can I infer about crop yield for a future season?”You get per-field vegetation trajectories across the growing season, the input a yield model needs, by mosaicking Sentinel-2 imagery over time and summarizing it inside detected field boundaries.

Building Sentinel-2 mosaics

Assemble cloud-free imagery across an area and a season.

Detecting field boundaries

Delineate every field so metrics can be summarized per parcel.
“How much tree canopy do we have, and where is it declining?”You get a canopy height surface plus per-parcel and per-zone summaries, by estimating height with RasterFlow and aggregating it with zonal statistics.

Estimating canopy height

Predict tree heights from aerial imagery.

Zonal statistics

Summarize raster values inside your own boundaries.
“How much of our property portfolio sits within a flood zone and how much asset value is exposed?”You get exposed parcel counts and assessed value at risk, broken out by city and hazard zone, by spatial-joining policy or parcel records against FEMA hazard polygons in WherobotsDB.

California coastal flood risk

Quantify parcels and assessed value inside FEMA flood zones.
“Are our risks quietly concentrating somewhere dangerous?”You get statistically significant hotspots, by applying Getis-Ord Gi* to claims, losses, or insured values across your book.

Hotspot detection

Separate real concentrations from noise with Getis-Ord Gi*.
“What is this watershed made of?”You get land cover composition per watershed, district, or custom boundary, by summarizing ESA WorldCover’s 10 m land cover classes by zone.

Exploring ESA WorldCover

Explore global 10 m land cover classes from ESA WorldCover.

Zonal statistics

Roll land cover up to the boundaries you care about.
“How do we work with satellite imagery without downloading any of it?”You get to query, filter, and process imagery collections in place, by reading STAC-compliant catalogs directly into WherobotsDB.

STAC reader and API

Search and load imagery collections without moving files.
“Are existing sidewalks viable pathways for autonomous or human delivery?”You get the pedestrian network as routable geometry, with gaps and discontinuities visible, so you can judge which segments carry a delivery robot or a courier on foot.

Detecting sidewalks with RasterFlow

Turn detected pathways into vector geometry you can route on.
“How do we keep our analysis from going stale without re-running it by hand?”You get a scheduled pipeline that refreshes the result on its own cadence and alerts you when it fails, by orchestrating the notebook or SQL job from Airflow.

Wherobots Airflow provider

Run notebooks, SQL, and job runs on a schedule.
Each page below maps an industry’s workflows to the Wherobots features and runnable example notebooks that support them.

Land Monitoring & Catastrophe Prediction

Land change detection, hazard exposure modeling, compliance checks, and vegetation analysis across large footprints.

Facility Management

Roof and site feature detection, severe weather exposure, nearby services, and infrastructure detection across distributed properties.

Site Selection

Trade area analysis, demographic scoring, competitor proximity, and pedestrian access for location decisions.

Fleet Logistics

GPS map matching, drive-time service areas, outlier detection, and road and sidewalk detection.

Agriculture & Forestry

Field boundary detection, growing season imagery, canopy height estimation, and per-field zonal statistics.

Insurance & Risk

Flood exposure analysis, underwriting enrichment, severe weather data, and portfolio concentration monitoring.

Environmental Monitoring

Change detection, severe weather data, land cover analysis, and imagery from STAC catalogs.

Last Mile Delivery

Pedestrian network mapping, delivery zone modeling, drop-off resolution, and coverage maps for the final leg of the delivery journey.

Ready to get started?

Start a free trial and run your first spatial analysis in minutes.