How the scoring works, and what we refuse to do

Deterministic, versioned scoring

Every ranking on this site comes from a versioned scoring function — currently v1.0-commute — applied to structured, sourced data. Hard constraints (your commute ceiling) eliminate candidates first; the remaining cities are scored on weighted fit, normalized 0–100. The same inputs always produce the same ranking, and every result shows the reasons it gained or lost points. Nothing editorial, nothing hidden, no model quietly choosing winners.

What the score covers today

Version v1.0-commute scores commute fit, on a 60-route grid computed via the OSRM v5 routing model (free-flow driving — typical minutes without traffic). Price bands, housing stock, and new-construction dimensions are displayed as pending and carry no weight until county-verified data is loaded. We would rather ship an honest one-dimension score than a five-dimension score with invented numbers.

The rule we won't break

No figure is ever fabricated to complete a comparison. When data is missing or stale, the display says so and the confidence badge drops. Every material number carries a source and retrieval date, and high-volatility data expires sooner than structural facts. Corrections are published openly on the about page.

Fair housing

Scoring uses objective, user-selected criteria only: workplace location, commute tolerance, budget, and housing requirements. We do not ask for, infer, or score on demographic characteristics, and we do not characterize any community by who lives there. Equal Housing Opportunity.

Source register