GeoJSON

What it is

The coverage grid, exported as a standard FeatureCollection that loads directly into mapping tools without transformation.

One Point feature per grid point, with the coverage statistics as properties.

A complete feature


{
  "type": "Feature",
  "geometry": {
    "type": "Point",
    "coordinates": [-180.0, -90.0]
  },
  "properties": {
    "coverage_percent": 0.0,
    "max_fold": 0,
    "max_gap_seconds": 0.0,
    "mean_fold": 0.0,
    "mean_gap_seconds": 0.0,
    "outage_count": 0
  }
}

Real output. The example is the South Pole, uncovered by a 53-degree constellation.

Coordinates are longitude first

GeoJSON coordinates are [longitude, latitude]. Every other surface in this software lists latitude first.

[-180.0, -90.0] is longitude -180, latitude -90: the South Pole. Read the other way it is latitude -180, which is not a valid latitude at all, and a tool that accepts it will place your data somewhere meaningless.

The ordering is mandated by the GeoJSON specification, not chosen here. It is the standard cause of a map with everything transposed.

Properties

Parameter	Type	Unit	Default	Required	Description
`coverage_percent`	number	percent	—	No	Fraction of the window this point was covered.
`max_fold`	integer	count	—	No	Most satellites visible at once from this point.
`mean_fold`	number	count	—	No	Average number visible.
`max_gap_seconds`	number	s	—	No	Longest interval with no satellite visible.
`mean_gap_seconds`	number	s	—	No	Average gap length.
`outage_count`	integer	count	—	No	Number of distinct gaps.

The export carries more per point than the coverage response prints. max_fold, mean_fold, mean_gap_seconds, and outage_count appear here and not in the console summary.

If you want per-point redundancy or outage counts rather than aggregate statistics, this is where they are.

Producing it

The exporter takes the analysis object that a coverage run returns, not a constellation identifier. It is a second step:


# 1. Run the analysis.
curl -s -X POST http://127.0.0.1:8080/coverage \
  -H 'content-type: application/json' \
  -d '{"constellation_id":"gj","duration_hours":1.0,
       "step_seconds":600.0,"grid_deg":45.0}' > cov.json
 
# 2. Wrap its `analysis` object and export.
python3 -c "import json;print(json.dumps({'analysis':json.load(open('cov.json'))['analysis']}))" \
  | curl -s -X POST http://127.0.0.1:8080/export/coverage/geojson \
      -H 'content-type: application/json' --data-binary @- > coverage.geojson

Passing a constellation_id instead fails:


Failed to deserialize the JSON body into the target type: missing field `analysis`

Points, not polygons

Each feature is a Point at a grid node, not a polygon covering the cell it represents.

That matters for rendering. A point layer at 45-degree spacing draws 40 dots, not a filled map, and most mapping tools will need either a symbol size that implies the cell or an interpolation step to produce a surface.

It also matters for interpretation: the statistic belongs to the node, and nothing is claimed about the area between nodes. A hole smaller than the grid spacing falls between points and is invisible.

Grid size and file size

The grid is two-dimensional, so the feature count grows with the square of resolution. At 45 degrees the example produced 40 features; at 5 degrees the same world grid produces 2664, and at 2 degrees roughly 16,000.

A fine grid is what you want for a quoted result and is often more than a browser-based map will render smoothly. Export coarse for a picture and fine for a number.