Skip to content
Home Analytics Case Studies Connect Start a conversation
Urban Analytics

A method chosen to fit the question — not the other way around.

Urban Analytics is how we combine machine learning, econometrics and geospatial data into one workflow: understand the problem, model it honestly, and deliver an answer that holds up.

Cluster analysis

Let the market tell you how it groups itself.

The same properties, first as an undifferentiated pile — then as the segments buyers actually price by.

1
MARKET CLUSTERS

Raw sales — one undifferentiated market, priced by averages.

Scroll to segment ↓

Cluster
Unsupervised learning surfaces the segments hidden in the data — no assumptions imposed up front.
Ensemble
Many models vote. The ensemble is steadier and more accurate than any single estimator alone.
Validate
Held-out sales test every estimate, so accuracy is measured on data the model never saw.
Econometrics & GIS

Isolate the effect. Quantify the dollars.

When the question is causal — what does this thing actually do to value? — averages aren't enough. Econometric models hold everything else constant and let the effect stand alone, mapped precisely in space with our integrated GIS.

Every sale
Price, structure, location and time — controlled together.
The effect alone
A single, defensible coefficient — in real dollars.
1,000+
home sales analyzed in a single pipeline effect study — enough to measure a proximity effect with statistical confidence.
Built to be challenged
Every specification, control and assumption is documented — because a finding that can't withstand cross-examination isn't worth much.
Geospatial & LIDAR

The part of value a spreadsheet can't see.

Proximity, elevation, sightlines and noise are physical facts. Our integrated GIS and remote-sensing LIDAR turn them into precise variables the model can use.

LIDAR point cloud
Spatial variables
PER PROPERTY
Distance & proximity
pipeline · highway · amenity
Elevation & noise
LIDAR-derived surfaces
Model input — a hard, comparable number
One of the first large-scale uses of remote-sensing LIDAR to value a community disamenity — highway elevation and traffic noise.

See the method on your data.

Tell us the property question. We'll show you which of these methods fits — and what the evidence can support.

Start a conversation