depth
Every place becomes a high-resolution intelligence node.
Evendo does not stop at mapping places. Each location is enriched with semantic attributes, experiential vectors, spatial context, behavioural signals and execution-relevant state.
Canonical Place Node
One place. Hundreds of signals. Structured for AI-native reasoning.
150+semantic attributes
65+experiential vectors
2M+mapped place nodes
Semantic density
Not listings. Not metadata. Structured real-world understanding.
The LocationGraph is designed to represent the qualities that matter when AI systems need to reason about where to go, what fits, what is nearby, what is practical, and what can be executed.
Node anatomy
A place is represented through layered meaning.
Every node combines canonical identity with semantic, spatial, experiential and execution-aware signals. The result is a place graph AI can reason over — not a directory it can merely search.
Canonical identity
Name, type, hierarchy, location, containment and disambiguation.
Semantic attributes
Structured traits describing ambience, suitability, access, locality and operational context.
Experiential vectors
Quantified signals for intent matching beyond category search.
Execution-relevant state
Signals that help bridge discovery into real-world coordination and action.
65+ experiential vectors
The graph captures qualities humans normally describe in language.
Atmosphere, audience fit, sensory profile, practical accessibility, temporal context and local character become structured signals AI can compare, rank and reason over.
Who a place works for.
Evendo models suitability across traveller types, social dynamics, group context and behavioural fit.
Family friendlySolo fitCouplesLocal crowd
What a place feels like.
Instead of relying only on category labels, the graph represents atmosphere and emotional texture.
QuietLivelyRomanticIntellectual
Whether a place is realistic.
Real-world discovery depends on transport, accessibility, timing, cost and effort — not just relevance.
Public transportWalkableAccessibleWeather fit
How the environment presents.
Architectural, visual and spatial qualities become queryable dimensions for AI-native discovery.
HistoricModernScenicDesign-led
What kind of intent it satisfies.
Places are mapped against the purpose behind a visit, not just the object category.
CulturalRelaxingEducationalSocial
How embedded it is in place.
The graph distinguishes tourist-heavy, local, hidden, mainstream, premium and culturally specific signals.
LocalHidden gemTourist-lightDistinctive
Beyond category search
Depth turns vague intent into structured reasoning.
“A good restaurant” is not a useful instruction. “Quiet, design-led, walkable, low-tourist, suitable for a solo evening near Shoreditch” is real-world intent. The LocationGraph is built to understand that level of detail.
01Human intent becomes semantic constraints.
02Constraints map onto attributes, vectors and spatial relationships.
03Places are ranked by experiential fit, not keyword overlap.
04Discovery can transition into real-world action.
Completeness by dimension
The dataset is deep because the real world is multi-dimensional.
Useful AI-native place intelligence must know more than location and category. It needs enough structured context to make judgments that feel human — and execute them reliably.
What depth enables
More dimensions create more precise real-world answers.
| Question type | Flat listing answer | LocationGraph answer |
|---|---|---|
| Category | Restaurants in London | Restaurants with the right atmosphere, audience fit, access pattern and local context |
| Intent | Popular attractions | Quiet, intellectually stimulating rainy-day cultural options avoiding tourist-heavy environments |
| Planning | Nearby places | Spatially coherent sequences shaped by distance, timing, mood and suitability |
| Execution | Bookable inventory | Real-world places with enough context to coordinate action, even across fragmented supply |
Compounding depth
The graph gets stronger as the world is used.
Semantic depth is not a static dataset. It compounds through coverage expansion, behavioural feedback, structured demand, verification and execution outcomes.
Coverage adds surface area
More mapped places increase the number of relationships the graph can understand.
Depth improves reasoning
More attributes and vectors improve the system’s ability to match nuanced intent.
Execution validates reality
Real-world outcomes reinforce which signals matter and where the graph needs refinement.
LocationGraph depth
Structured understanding of the physical world.
Evendo turns places into rich, persistent, AI-readable entities — enabling discovery systems and agents to reason with the nuance real-world intent requires.