City Twins
Loved a city? Find 10 more like it.
Why it exists
Every where-should-I-live tool is a quiz that asks people to rate walkability on a 1 to 5 scale. Nobody can answer that honestly. RVers already know a city that felt right, and no tool started from there.
What I built
A matching engine that starts from the city you loved: weighted similarity across vibe scores, size, climate, cost, and demographics, with priority chips that re-rank live around what mattered to you (mountains, college energy, dry air). Hard filters and a sortable 16-column table serve the spreadsheet crowd. The whole dataset is static JSON baked at build time. No database, no accounts.
What it does
- 471 cities with sourced data: Census ACS 2024 demographics, ERA5 climate actuals, NOAA snowfall normals, every pipeline keyless and $0 to run
- Idea to deployed prototype in one day; real data and 140 small RV towns (Moab, Sedona, Bar Harbor) landed over the next two
- The matches pass the lived-there test: Fort Collins returns Bozeman, Boulder, Missoula. Moab returns Kanab and Springdale
What broke
NOAA encodes zero-to-trace snowfall as -7777. Sixteen Gulf towns got -778 inches of snow, log1p returned NaN, and a single NaN in a sort comparator silently scrambled every ranking on the site while the scores still looked right. One clamp fixed it. A validation guard that fails the build on any non-finite field keeps it fixed.
Want something like this built for your team?
Start a project