Geo Filters

If that feature has been activated on the server, Castellum also allows to filter by geographical properties of a subject’s address.

Example

If the desired participants for a study should be older than 40 years and live in an area where the noise from an airport is over 40dB you would create and upload a GeoJSON file with a (multi)polygon spanning the area where this condition applies. Afterwards proceed as usual with setting the filter for age in the study filter UI. The resulting subjects in the recruitment interface match the filter condition AND their contact address is located inside the uploaded polygon.

GeoJSON format

To use this, you need to provide a valid GeoJSON file with a either a single feature or a feature collection of type “Polygon” or “MultiPolygon”. Castellum will then filter all subjects that live inside this (multi)polygon, so it should enclose all areas where the conditions for recruitment are met. Note that the Coordinate Reference System (CRS) must be WGS84 (EPSG:4326).

This is an example of a valid GeoJSON file:

{
  "type": "Feature",
  "properties": {},
  "geometry": {
    "type": "Polygon",
    "coordinates": [[
      [13.3517429, 52.5189382],
      [13.2908489, 52.5165476],
      [13.3463410, 52.4848593],
      [13.3512518, 52.4468630],
      [13.3964313, 52.4785787],
      [13.4627272, 52.4504546],
      [13.4308069, 52.4893448],
      [13.4730399, 52.5192370],
      [13.4141102, 52.5180417],
      [13.3821899, 52.5604549],
      [13.3517429, 52.5189382]
    ]]
  }
}

http://geojson.io is a simple tool to preview and edit GeoJSON.

How to convert a shapefile to GeoJSON

Most Geographic Information Systems (GIS) such as QGIS or ArcGIS support the conversion to GeoJSON out of the box. Desktop GIS applications should automatically reproject the shapefile to WGS84 during the export to GeoJSON format.

Optimize GeoJSON for performance

The duration of the calculation scales roughly linearly to the number of vertices in the GeoJSON file. If the calculation takes too long, try to reduce the number of vertices in the file. This can be achieved in two ways:

Crop the spatial extent of your geographic data to the extent of the area where the subjects live (or at least most of them). This makes sense if most of your subjects are from the same city or state.
Generalize your polygon. This reduces the amount of vertices while retaining the general shape of the polygon. This makes sense of the resolution of your data is very high. Remember that Castellum derives the subjects’ geolocations from their addresses, which refer to houses. Therefore, a polygon with a resolution of 10cm has no advantage over one with a resolution of 10m, as address data is resolved to 10m at best, as determined by typical house sizes.