Methodology
How We Score European City Matches
We built a system to answer one question: Where in North America feels most like Europe?
Every city is scored across three dimensions. Vision is weighted highest (40%) because visual character is the strongest driver of perceived "European feel" in user testing. The final match percentage is a weighted combination of all three scores.
The Formula
Total Match = Vision × 0.40 + Topology × 0.30 + Amenity × 0.30
Each dimension is scored 0–10. The total match percentage is the weighted average × 10.
Example: Why Jacksonville Beach Scores 77% for Biarritz
7.4
Vision AI
× 40%
7.6
Topology
× 30%
8.2
Amenity
× 30%
7.4 × 0.40 + 7.6 × 0.30 + 8.2 × 0.30 = 7.7 → 77% match
What We Measure
Vision AI
40% weightWe sample 50–200 street-level images per city and compare them against a reference dataset of European cities (Paris, Barcelona, Rome, Amsterdam, etc.). The model evaluates building height-to-street-width ratios, facade density, roofline patterns, and architectural ornamentation. Output: 0–10 score.
Metrics
- Building height / street width ratio
- Facade ornamentation density
- Roofline patterns (mansard, gabled, flat)
- Street-level enclosure ratio
Example: Jacksonville Beach: 7.4 — pier + beachfront density resembles Biarritz
Street Topology
30% weightWe analyze the street network within 2km of each city center using OpenStreetMap data. European cities typically have higher intersection density (intersections per km²), shorter average block lengths, and more organic (less grid-like) street patterns. We measure curvature variance and block size distribution.
Metrics
- Intersection density (per km²)
- Average block length (meters)
- Street curvature variance
- Grid regularity index
Example: Jacksonville Beach: 7.6 — coastal grid + walkable core pattern
Amenity Density
30% weightWe count walkable amenities within 1km of city center: cafés, restaurants, bars, parks, museums, transit stops, and retail. European cities consistently show higher density of mixed-use destinations. Data from Geoapify API and OpenStreetMap.
Metrics
- Cafés + restaurants per km²
- Parks and public spaces (count)
- Transit stops within 1km
- Mixed-use zoning ratio
Example: Jacksonville Beach: 8.2 — high café + restaurant density along beachfront
Confidence Levels
We report confidence based on score variance across dimensions:
High Confidence: All 3 dimensions within 1.5 points (strong agreement)
Medium Confidence: Dimensions vary by 1.5–3 points (mixed signals)
Low Confidence: Dimensions vary by 3+ points (sparse or noisy data)
What We Don't Measure (Yet)
Our model focuses on measurable, physical characteristics. We do NOT currently capture:
- Cultural identity, language, or local customs
- Historical depth (age of buildings, centuries of history)
- Subjective "feel" beyond physical form
- Safety, cleanliness, or tourist infrastructure
A score of 80% does NOT mean a city is "80% European." It means the city ranks similarly across measurable features compared to its European twin.
Quick Reference
High Topology Score =
Tight streets, irregular grid, organic growth patterns (like European medieval cores)
Low Topology Score =
Wide roads, rigid grid, suburban sprawl patterns (typical American planning)
Scores like "7.4" or "8.2" should be read as relative rankings, not absolute measurements. A 7.4 Vision score doesn't mean "74% European" — it means this city's visual character ranks similarly to other cities in the 70–79% range.
These are carefully calibrated heuristics — useful for comparison and discovery, not for defining a city's entire character. Use this system to find interesting places, then explore them on your own terms.