Given a geojson of spaces on a floor, this endpoint returns person-count occupancy distribution histograms per normalized space type (NST), grouped by capacity. For each combination of NST and capacity, the response describes what percentage of 1,000 Monte Carlo simulation runs resulted in each possible occupancy level (0 to capacity), pooled across all spaces sharing that NST and capacity.

Request Body

The supplied input is expected to be valid GeoJSON. At the top level, there is expected to be a person_count attribute whose value has the following specification:

(int, required) person count value to model

Within the GeoJSON specification, this type of attribute is considered a "foreign member". See the Predict API GeoJSON section below for more information regarding the input expectations.

Response

The response will conform to the JSON API spec. The top-level keys will be data and meta.

data will contain as many objects as person_counts values requested. Within each object the attributes will be person_count and normalized_space_types.

person_count: The person count simulated.
normalized_space_types: An object keyed by snake_case NST name (e.g. enclosed_collab). Each value is a flat array of capacity-group objects, one per distinct capacity found among spaces of that NST.
- capacity: (int) The capacity shared by all spaces in this group.
- spaces: An array of the constituent spaces in this group, each with:
  - id: The original space ID supplied through the GeoJSON.
  - name: The original space name supplied through the GeoJSON.
- histogram: An array of objects covering occupancy levels 0 through capacity (inclusive). Each entry has:
  - person_count: (int) The occupancy level (0..capacity).
  - pct: (float) The percentage of simulation instances in which this occupancy level occurred across all spaces in the group. Values sum to 100 within each capacity group.

The meta object includes:

person_counts: The array of requested person counts.
n_runs: (int) The number of Monte Carlo simulation runs performed (1000).
model_version: The model version string.