`elwood_spatial.network`

Spatial network construction from GeoDataFrames. Builds distance-band neighbor graphs using libpysal, where each device is connected to others within a threshold distance.

Sensor networks at 5km, 7.5km, 12.5km, and 20km distance thresholds in the Methow Valley, WA

Network Type

Network = dict[str, dict[str, list]]
# Format: {device_id: {"neighbors": [...], "weights": [...]}}

`build_network(gdf, threshold, id_column, ...)`

Build a spatial network from a GeoDataFrame of device locations.

Parameter	Type	Default	Description
`gdf`	`gpd.GeoDataFrame`		Device locations with point geometry
`threshold`	`float`		Distance threshold for neighbor inclusion
`id_column`	`str`		Column containing device identifiers
`max_neighbors`	`int`	`20`	Maximum neighbors per device
`min_neighbors`	`int`	`3`	Minimum neighbors required
`projected_crs`	`str \| int \| None`	`None`	CRS to project to before distance calc

Returns Network.

import geopandas as gpd
from elwood_spatial.network import build_network

gdf = gpd.read_file("sensors.geojson")
network = build_network(gdf, threshold=50_000, id_column="sensor_id")

`filter_network(network, active_ids, min_neighbors=3)`

Filter a network to only include active devices, removing any that fall below the minimum neighbor count.

Parameter	Type	Default	Description
`network`	`Network`		Full network
`active_ids`	`set[str] \| list[str]`		IDs of currently active devices
`min_neighbors`	`int`	`3`	Minimum neighbors to retain a device

Returns Network.

`get_neighbors(network, device_id)`

Get the neighbor list for a device.

Parameter	Type	Description
`network`	`Network`	Spatial network
`device_id`	`str`	Target device

Returns list[str]. Empty list if device not found.