4.8M property nodes connected into a single valuation pipeline

Twelve feeds described the same buildings with four different identifier schemes. Reconciling them was the whole project; the model was the easy half.

NDAData & Platform Engineering

4.8M
property nodes reconciled: 12
source feeds unified: 6x
faster valuation runs

Abstract computational geometry representing distributed balanced systems in clean architectural line work.

The constraint

Twelve data feeds described the same physical buildings using four incompatible identifier schemes. The valuation model was straightforward; knowing which records referred to the same building was not.

What we changed

We built the entity-resolution layer first and treated it as the product, with the valuation pipeline as a consumer of it rather than the other way around.

Geospatial matching as the tiebreak when identifiers disagree
Confidence scores surfaced rather than hidden
Manual adjudication queue for the genuinely ambiguous

Where it landed

4.8 million property nodes reconciled into one graph, and valuation runs six times faster because they no longer redo the matching each time.

Have a system to ship

Bring us the problem that keeps stalling.

A 30-minute working session with the engineers who would do the build — not an account manager. You leave with an architecture opinion either way.

Book a call Send a message

Know a market better than we do

Partner with us on your domain.

You bring the domain judgement and the relationships. We bring the engineers who turn it into shipped software. Revenue-shared, no exclusivity.

How partnership works