How this works

The dimensions

Every neighborhood is scored on these 18 dimensions. Each score lives in [-1, 1]. 8 are symmetric (both poles are real lived preferences — landing on the wrong one hurts in either direction). 10 are asymmetric needs(only the high end is an active preference; landing low just means “not a driver” and the neighborhood’s abundance can’t hurt you).

• Urban Intensity Tolerance

  Symmetric

  How well you thrive in dense, stimulating, high-energy environments. People high on this dimension feel energized by crowds, noise, nightlife, and unpredictability. People low on it need calmer streets and quieter blocks to function. Both poles are real preferences.

  Low · Drained by intensity. Needs calm. · High · Energized by density and stimulation.

• Transit Psychology

  Asymmetric need

  How much you value transit redundancy and optionality. People high on this dimension feel constrained by single-line dependence. People low on it are fine with one good line. Not needing redundancy doesn't mean you'd dislike a transit-rich neighborhood. The penalty applies only when you want optionality and the neighborhood lacks it.

  Low · Tolerant of single-line dependence. · High · Needs transit optionality and redundancy.

• Resident Wealth / Market Tier

  Symmetric

  How much you weight neighborhood resident-wealth tier and aspirational market positioning. Both poles are real preferences. The high pole is established-wealth, high-market-tier, status-address territory (UES, Greenwich, Tribeca). The low pole is working/middle-income, value-coded territory (Bushwick, Sunset Park, Inwood). Picking value-over-stretch is an active preference, not neutrality. Distinct from social-register (polish/aesthetic) and amenity-function (real-quality substance). A neighborhood can be wealth-tier-high but amenity-function-mid (UWS), or wealth-tier-low but amenity-function-high (Flushing).

  Low · Working/middle-income, value-coded. · High · Established wealth, high market tier.

• Space Sensitivity

  Symmetric

  How much physical space at home affects your wellbeing. Both poles are real lived preferences. Picking "I can be happy in a small space" is also picking against neighborhoods of huge apartments far from the action — a real tradeoff, not just non-preference.

  Low · Tolerates small spaces gracefully. · High · Needs spaciousness to feel at home.

• Future Family Orientation

  Asymmetric need

  Whether kids are part of your household or near-term plan. People high on this dimension want family-life infrastructure (schools, playgrounds, kid-friendly streets, other families). People low on it just don't need that — it does not mean they actively dislike family-coded neighborhoods. The penalty applies only when you want family infrastructure and the neighborhood lacks it.

  Low · Not a factor in this window. · High · Plans for or has kids in the household.

• Cultural Ecosystem Alignment

  Asymmetric need

  How important culturally familiar food, immigrant communities, language access, and cultural rhythm are to your daily life. Not needing cultural anchor doesn't mean you'd dislike a culturally rich neighborhood. Penalty applies only when you want this and the neighborhood is culturally neutral.

  Low · Neutral on cultural ecosystem. · High · Needs cultural familiarity in daily life.

• Environmental Openness

  Asymmetric need

  How much access to parks, waterfronts, visual openness, and walking paths matters to your emotional baseline. Not needing nature doesn't mean you'd dislike a green neighborhood. Penalty applies only when you need open space and the neighborhood lacks it.

  Low · Content in fully built environments. · High · Needs nature, parks, or open vistas.

• Creative Scene Density

  Symmetric

  How much active art, indie music, and maker culture lives in the neighborhood — galleries, music venues, studios spilling onto sidewalks, the kind of culture you stumble into walking around. Both poles are real lived preferences. Some people genuinely want this density at home. Others want calm distance from it.

  Low · Prefers calm distance from the creative scene. · High · Wants the creative scene at the doorstep.

• Daily Friction Sensitivity

  Symmetric

  Both poles are real lived preferences. Some people are energized by noise, density, and friction. Others are drained by it. Mismatching either way is a real friction. (Neighborhoods are scored such that high values mean low friction — calm, orderly.)

  Low · High tolerance for daily chaos. · High · Drained by noise, disorder, friction.

• Perceived Safety Need

  Asymmetric need

  How much you want to feel safe walking home alone, late at night, or with kids. This is a perception measure, not a crime-statistics measure. Most NYC neighborhoods are statistically safe by national standards, but perception varies. Higher need = stronger filter against neighborhoods with rougher reputations. The dimension is asymmetric: low need doesn't mean disliking safe places, just not weighting it heavily.

  Low · Not a strong factor for me. · High · I need to feel safe at all hours.

• Zoned School Quality

  Asymmetric need

  How much zoned public school quality matters to you. Most relevant to people with kids or planning to have them, but also affects long-term resale and neighborhood stability for everyone. Asymmetric: not caring about schools doesn't mean disliking neighborhoods that happen to have good ones.

  Low · Not a factor for me. · High · Top-rated schools required.

• Social Register

  Symmetric

  Cultural code — how formal-traditional vs. progressive-casual the dominant social register is. NOT visual polish (handled by streetscape-quality), NOT resident wealth (prestige-orientation), NOT substance/quality (amenity-function). Formal-traditional means doorman-fluent, status-through-restraint, traditional dress codes — UES, Greenwich CT, Carnegie Hill. Progressive-casual means neighborly-informal, intellectual, values-signaling — Park Slope, Williamsburg, parts of UWS. Park Slope can be wealthy AND have polished streetscape AND have real substance AND still register as progressive-casual on this axis. Both poles are real lived preferences.

  Low · Progressive, casual, intellectual, neighborly-informal. · High · Formal, traditional, doorman-fluent, status-through-restraint.

• Visitor-Facing Energy

  Symmetric

  Whether the public realm is for residents or for visitors and destination-seekers. Resident-rooted streets are routine-driven, with locals walking dogs and ground-floor retail serving residents. Destination-facing streets are stages: brand retail, weekend crowds, recognizable spots people travel cross-town to reach. Some users want iconic visitor energy; others actively dislike living inside someone else's itinerary.

  Low · Resident-rooted, routine-driven, locals only. · High · Destination-facing, iconic, visitor-heavy.

• Rootedness vs Access

  Symmetric

  Whether the neighborhood rewards being rooted in a local community (knowing your coffee shop owner, your block, neighbors who recognize you) or rewards having access to everything (walking distance to world-class restaurants, citywide cultural density, late-night anything). The closest psychographic proxy for borough identity: rootedness-leaning self-selects into Brooklyn brownstone belt, Queens enclaves, and suburbs; access-leaning self-selects into downtown Manhattan, Midtown, and UWS/UES corridors. Both poles real lived preferences.

  Low · Rooted. Local community, neighborhood character, knowing the regulars. · High · Access. Everything-at-your-fingertips density, world-class options walking-distance.

• Daily-Life Walkability

  Asymmetric need

  How easily a resident can walk to daily-life infrastructure: groceries, gyms and fitness studios, pharmacies, dry cleaners, cafés. Distinct from rootedness-vs-access (which is about destination/cultural density) — this is about whether daily errands and routine wellness happen on foot. Critical for suburb discrimination (Larchmont walkable downtown vs Cresskill car-only), and surfaces a real gap in some NYC core neighborhoods (Hudson Yards has limited groceries despite density). Asymmetric: not needing walkable amenities doesn't mean disliking neighborhoods that have them.

  Low · Not a daily concern. I'll drive or travel for these. · High · Essential. I want groceries, gyms, errands within easy walking distance.

• Community Fabric

  Asymmetric need

  Whether community life is organized around public village overlap or private estate seclusion. Civic-village neighborhoods stitch families together through public schools as the social hub, town events, libraries, walkable downtowns, and broad civic engagement, so residents keep bumping into each other. Estate-and-club neighborhoods stitch life together through country clubs, private schools, larger lots, and curated by-invitation circles, with less daily mixing on shared public ground. Both are durable lived preferences. The axis that splits Maplewood and Larchmont from Scarsdale and Greenwich, even when prestige tier and school quality match.

  Low · Estate, club, private-school, by-invitation social circles. · High · Public-civic, walkable downtown, public-school as community hub.

• Streetscape Quality

  Asymmetric need

  Block-to-block walking pleasure — tree-lined sidewalks, prewar facades, stoops and storefronts, the visual-and-pedestrian texture of the neighborhood itself. Distinct from daily-life-walkability (errand reach), environmental-openness (parks/waterfront/visual openness), and rootedness-vs-access (cultural density). This dim is *only* about whether the streets themselves are pleasant to walk on. Low means absence of preference, not active dislike.

  Low · Functional streets are enough. · High · Wants streets that are pleasant to walk for their own sake.

• Amenity Function

  Asymmetric need

  Real walkable-distance access to substance-over-presentation amenities: food, groceries, cafes, retail that genuinely deliver real quality and value. Distinct from streetscape-quality (the look of streets), social-register (polish/aesthetic), and prestige-orientation (resident wealth tier). High score means real quality regardless of polish. Flushing-tier real-deal grocery counts; Hudson Yards mall does not. Asymmetric: not needing high-function amenity doesn't mean disliking neighborhoods that have it. Penalty applies only when user wants substance and neighborhood is thin.

  Low · Basic amenities are enough. · High · Needs food, groceries, cafes, retail that genuinely deliver real quality and value.

The questions

Across the 16 questions, three formats appear:

• Forced-choice (5). Two or three mutually exclusive options. Each option contributes a small impact to one or more dimensions.
• Slider (3). A 5-point agree/disagree scale on a single declarative statement. Snaps to -1, -0.5, 0, 0.5, 1.
• Multi-select (8). Checkboxes. Used for cultural-community tags and non-negotiable filters.

The vector of dimension scores is derived from your answers, not accumulated. Editing an earlier answer recomputes everything from scratch — no drift across back-navigation.

The math

For each neighborhood, the engine measures how far its profile sits from yours, dimension by dimension. Symmetric dimensions contribute the squared difference in either direction. Asymmetric-need dimensions contribute the squared shortfallonly — if the neighborhood over-delivers what you’d already deemed unimportant, no penalty. Distances are mapped to a 0–100% match score.

A small cultural-tag boost (+8% per match, capped) rewards alignment between cultural communities you selected and the neighborhood’s tagged communities. Non-negotiable must-haves operate as hard filters — neighborhoods that fail any one are excluded from the ranked list entirely, with the failing constraint surfaced so you can see why.

Where the data comes from

• Editorial scoring on most dimensions (urban-intensity, social-register, visitor-facing-energy, daily-life-walkability, etc.) is the result of a dual-model pass: two large language models score every neighborhood independently at maximum reasoning depth, the scores are averaged within a tight disagreement bound, and edge cases are reconciled by hand.
• School-quality scores use NY State DOE data plus Niche as a cross-reference, normalized to the same [-1, 1] scale.
• Commute minutes are computed door-to-door via the Google Routes API for nine office clusters across NYC, NJ, and Connecticut. Editorial overrides patch a few cases where automated routing misreads available transit (e.g., the Metro-North New Canaan branch).
• Cultural tags, housing types, car-dependence flags are hand-curated from on-the-ground knowledge of the neighborhoods.
• Neighborhood photographs (113 of 115) are sourced from Wikipedia’s lead images, with attribution and license shown on each neighborhood page. The remaining neighborhood falls back to an illustrated borough hero.

Archetypes

Each result page opens with a named archetype: a 8-way clustering of common lifestyle profiles in this metro. The match is the archetype whose profile sits closest to your vector. The archetype label is descriptive, not prescriptive — your ranked neighborhoods come from your actual vector, not the archetype’s.

Honest limitations

• Coverage skews where users actually move.Manhattan, Brooklyn, Queens, NJ commuter belt, lower Westchester, southern CT, near-LI suburbs are well-mapped. The Bronx (Riverdale) and Staten Island (St. George) are intentionally lighter — most users moving into the metro don’t consider them.
• Suburbs share an engine designed for urban tradeoffs. Family-trajectory and school-quality dominate suburb scoring, which is roughly right but flattens differences among, say, Westchester villages. A separate suburb-tier calibration is on the roadmap.
• Cultural-community granularity is coarse.“East Asian” bundles Chinese, Korean, Japanese, and Taiwanese communities; “Latin American” bundles Mexican, Dominican, Puerto Rican, Colombian, and Ecuadorian. Splitting these is on the roadmap.
• Bundled neighborhoods get a single hero photo. Areas like Nolita / Little Italy and Flatiron / NoMad each show one Wikipedia photo of one half of the bundle.
• The map color gradient is stretched to your range.Worst score in your list renders red, best renders green, so the map stays readable even when raw scores cluster. Don’t read the colors as absolute fits.