VeloraMaps turns a location into a 0 to 100 Vibe Score by combining stable map context, nearby activity, road-noise proxies, amenities, nature, bounded weather context, and light calibration from user feedback. The model is not trying to declare a place objectively good or bad. It estimates how suitable a specific point feels for common intents such as staying, focusing, socialising, or relaxing, then shows the underlying pillar scores and confidence so the result can be questioned.
Each score starts with the exact coordinate selected on the map, then evaluates nearby signals around that point. Most map-derived data is cached server-side by rounded coordinate so repeat checks are fast and do not overload public map infrastructure.
Source: OpenStreetMap road geometry through the VeloraMaps OSM cache, plus time-of-day adjustment and optional user calibration. Refresh: OSM-derived signals are cached for roughly 12 to 24 hours; time adjustment recalculates on every visit.
Source: OpenStreetMap landuse, leisure, and natural features through Overpass or cached OSM signals. If OSM is temporarily unavailable, VeloraMaps can estimate from place category and name, labelled as approximate. Refresh: cached OSM data refreshes about daily.
Source: OpenStreetMap points of interest such as cafes, restaurants, shops, transport nodes, and social venues. Refresh: cached OSM signal results refresh on a 12 to 24 hour cycle, with live fallback when available.
Source: a blend of Quiet, amenities, bounded weather context from Open-Meteo, and nearby services from OSM. Refresh: weather updates live from Open-Meteo; map signals use the OSM cache; user calibration applies immediately on the device.
The interactive map uses formula version 2026-09-09.1. Raw signals become sub-scores, then the selected mode determines how they contribute. Stay prefers moderate activity, Focus and Relax prefer less activity, and Social prefers more. Very low quietness limits Focus and Relax; very low nature also limits Relax. Bonuses increase gradually rather than switching on at one threshold. Explicit search preferences retain their custom weighting.
Weather is held at neutral when calculating place fit, then adds a bounded adjustment of at most 12 points. Time and community adjustments, where shown, are included in the input scores. Missing evidence uses a labelled neutral 50/100 placeholder, not invented coordinate-based variation. Nature can instead use a labelled place-category estimate. Browser results are reused for up to ten minutes, or one minute when estimated, before a new request can retry them. This does not change the server's longer OSM cache lifetime.
Confidence describes source availability and age, not proven accuracy. Its indicative range is a heuristic, not a statistically validated prediction interval. Public city summaries use a separate four-pillar snapshot model; they are not a fresh calculation for every map click.
A low score usually means one or more core conditions are clearly working against the selected intent. Example: a busy road junction with little greenery may be low for Relax because traffic, noise, and limited nature all push against the goal.
A medium score means useful qualities exist, but with trade-offs. Example: a city centre street might be strong for Social because it has food and transport, but only medium for Stay because high activity can become tiring.
A high score means the local signals line up well for the selected use case. Example: Principe Real in Lisbon can score high because it combines gardens, cafes, walkability, and enough calm to feel balanced.
Vibe Scores are estimates, not guarantees. They do not measure personal safety, rent, crowd mood, air quality at street level, accessibility for a specific disability, opening hours, temporary roadworks, events, queues, or whether a venue is personally enjoyable. They also depend on public map coverage: some cities have richer OpenStreetMap data than others, and private spaces may be missing.
It captures nearby structure: road intensity, mapped green or natural areas, amenity density, activity proxies, live weather, and how those signals compare with the selected intent.
It does not replace local judgement. A high score should be treated as a useful shortlist signal, not a final decision about where to live, work, stay, or meet.