Methodology

BestCityFor is designed as a transparent decision engine. Every score is computed from public data, with explicit weights and “why it ranks” drivers.

Want the exact sources? Sources.

How the score is computed

  1. We select a set of metrics for each vertical (Remote Work, Retirement) and assign default weights.
  2. For each metric, we winsorize outliers (clamp to the 5th–95th percentile) and normalize onto a 0–100 scale.
  3. We apply weights and compute a weighted average. If a metric is missing for a city, its weight is removed and remaining weights re-normalize.
  4. We map the 0–100 score into a grade (A+ through F) and derive “drivers” from the biggest weighted contributions. Grades are assigned from score percentiles within the current city set (to avoid skewed distributions from metric outliers).

Remote Work weights

Defaults for the top 200 cities (ACS 2023).

MetricWeightDirection
Typical rent (2BR)30Lower is better
Broadband subscription15Higher is better
Work-from-home share15Higher is better
Median household income15Higher is better
Bachelor’s+ rate10Higher is better
Commute 45+ min share5Lower is better
Climate comfort10Higher is better

Retirement weights

Defaults for the top 200 cities (ACS 2023).

MetricWeightDirection
Typical rent (2BR)20Lower is better
Typical home value15Lower is better
Age 65+ share10Higher is better
Median household income10Higher is better
Broadband subscription5Higher is better
Commute 45+ min share5Lower is better
Climate comfort35Higher is better

Young Professionals weights

Defaults for the top 200 cities (ACS 2023).

MetricWeightDirection
Median household income25Higher is better
Bachelor’s+ rate15Higher is better
Work-from-home share10Higher is better
Broadband subscription5Higher is better
Typical rent (2BR)20Lower is better
Age 25–34 share15Higher is better
Below poverty line5Lower is better
Commute 45+ min share5Lower is better

Limitations

  • City-level averages can hide neighborhood differences.
  • Some important factors (tax policy, healthcare access, walkability) aren’t included in v1.
  • Data is updated on the ACS release cycle; NOAA may be included when available.

FAQ

Is this a subjective ranking?
It’s a model: we pick measurable proxies, normalize them across cities, apply explicit weights, and publish the components. You can disagree with the weights — that’s the point.
Why normalize instead of using raw values?
Raw metrics aren’t comparable across different units. Normalization maps each metric onto a common 0–100 scale so weights behave predictably.
What about neighborhoods or metro areas?
This is city-proper (incorporated places) and city-level averages. Neighborhood-level and metro-level choices can differ a lot.