The houses most likely to be a below market deal.

Every residential property in the city, scored on one thesis: the worst house on an improving block. Underpriced for its street, in a neighbourhood where owners are actively renovating. No 10,000 row list. Just the shortlist worth a phone call.

How it runs

In a live deployment it refreshes from city open data every run

1
Scan every property
Pulls every residential parcels from city open data. No MLS, no login, no vendor.
2
Score the gap
Compares each house to its neighbourhood median and the renovation activity around it. Cheap for its street plus a block that is upgrading means a deal.
3
Filter to real deals
Drops condos and brand new suburbs, keeps established blocks with a genuine mix of old and new. No noise.
4
Hand you the shortlist
Ranks what survives and surfaces the shortlist worth a call, mapped, with the reasoning on each.

Loading sample data

Where they are

lowerhigher score

What the score means

A number out of 100 answering one question: is this house cheap for its own street, on a street that is getting better? A 78 is a stronger candidate than a 62. It is a shortlisting signal, not a valuation.

How far under: the city's assessed value against what a typical house on that same street is assessed at. Roughly half the score.
Is the block improving: how many renovation permits the neighbours have pulled. Roughly the other half.
Has this house been left alone: no permits on file, no garage. A house nobody has touched while the street upgrades is the one worth a call.
Where we do not look: condos, and brand new suburbs where every house is the same age and there is no gap to find.

What the paid data layer adds

Year built and sale history (land titles), which removes new build noise
Owner tenure and absentee flags, a real motivation signal
Contact enrichment, which turns a list into outreach