hextraj

Hex-labelling of trajectory data.

Maps lon/lat positions to a projected hexagonal grid and provides tools for aggregation and connectivity analysis.

• hex_counts — heat maps and density aggregation

• hex_connectivity — origin-destination matrices

• hex_connectivity_power — multi-generation transport probabilities

• hex_connectivity_dask — lazy dask-native connectivity for large datasets

• Full dask support throughout

Installation

pip install git+https://github.com/willirath/hextraj.git@main

For dask, scipy, and cartopy support:

pip install "hextraj[full] @ git+https://github.com/willirath/hextraj.git@main"

Quick example

from hextraj import HexProj

hp = HexProj(lon_origin=-3.0, lat_origin=54.0, hex_size_meters=50_000)

# Label positions -> int64 hex IDs
hex_ids = hp.label(lon, lat)

# Build a GeoDataFrame with Polygon geometries
gdf = hp.to_geodataframe(hp.region_of_hexes(region_polygon))
gdf["count"] = counts.reindex(gdf.index).fillna(0)
gdf.plot(column="count", cmap="YlOrRd")

Notebooks

• Hex grid construction
  • Imports
  • Projection setup
  • 1. Rectangle of hexes
  • 2. Region of hexes from a shapely polygon
• Trajectory → hex aggregation → choropleth
  • Imports
  • Projection setup
  • Synthetic ship tracks
  • Label positions
  • Aggregate — count visits per hex
  • Build the full hex universe for the region
  • Join counts onto the hex universe
  • Plot choropleth
  • Hex outlines
  • Hex centers
  • Origin–destination edges
  • Combined plot: choropleth with OD edges overlay
• Hex connectivity — NW Shelf trajectories
  • Imports
  • Load trajectories
  • Set up HexProj
  • Label all positions
  • Dask recipe for larger-than-memory arrays
  • Visit heatmap
  • Initial and final position heatmaps
  • OD connectivity — efficient recipe
  • Combined plot
• Hex Analysis: Heatmaps and Connectivity
  • Setup
  • Load the NW Shelf trajectory data
  • Build a HexProj for the NW Shelf domain
  • Hex heatmap: hex_counts
  • Connectivity: hex_connectivity
  • INVALID_HEX_ID as a domain-exit signal
  • Normalisation: conditional transport probability
  • Multi-generation connectivity: hex_connectivity_power
• Dask-native OD connectivity with hex_connectivity_dask
  • Setup
  • Load data
  • Build HexProj
  • Build the lazy connectivity table
  • Understanding laziness
  • Filter before compute
  • Compute obs=0 → obs=19 connectivity
  • Visualise as a choropleth heatmap
  • Multi-obs view: spreading over time
  • Cumulative settling connectivity

API Reference

• API Reference
  • hextraj