Geofences

Geofences represent custom regions or places monitored in your project. Depending on your use case, a geofence might represent a retail store, a neighborhood, and so on.

Radar geofencing is more powerful than native iOS or Android geofencing, with cross-platform support for unlimited geofences, polygon geofences, isochrone (time-based) geofences, temporary geofences, and stop detection.

Geofences provides the following user context:

{
  geofences: [
    {
      tag: "store",
      externalId: "123",
      description: "Store #123",
      metadata: {
        parking: false
      }
    }
  ]
}

Geofences also provides the following events:

  • user.entered_geofence
  • user.exited_geofence

You can create geofences via the dashboard, CSV import, the API, nightly sync, or integrations.

Alternatively, if you don't have your own custom place data, Radar can detect when a user visits a place even if you haven't set up a geofence for that place. Learn more about Places.

How it works

Geofencing works in the foreground and in the background. All event generation happens server-side. This allows Radar geofencing to be more powerful than native iOS or Android geofencing, with cross-platform support for unlimited geofences, polygon geofences, isochrone (time-based) geofences, temporary geofences, and stop detection.

Radar generates a geofence entry event if a user enters a geofence (if stop detection is off) or stops in a geofence (if stop detection is on) with sufficient confidence, then a geofence exit event when the user leaves the geofence with sufficient confidence.

Create geofences

You can create geofences via the dashboard, CSV import, the API, nightly sync, or integrations. You can create geofences in the Test environment for development and testing, and in the Live environment for production.

You specify the metadata for geofences when you create them, including the tag (an optional group for the geofence, e.g., store), external ID (an optional external ID for the geofence that maps to your internal database, e.g., 123), and description (a display name for the geofence, e.g., Store #123).

Geofences should be uniquely referenced by tag and external ID, assigned by you when a geofence is created. To disable or update metadata for a geofence, re-import the geofence with the same tag and external ID.

Size and reliability

For maximum reliability, create geofences with a radius of 100 meters or bigger for circles or an area of 10,000 square meters or bigger for polygons, and add padding to compensate for location accuracy. While you can create geofences with a smaller size, events may be missed or delayed.

Accuracy and reliability may be further reduced if wi-fi is turned off, especially indoors, or if wi-fi is unavailable, such as in sparse suburban or rural areas. In general, these restrictions apply to all apps and SDKs using location services on iOS and Android, not just Radar.

Dashboard

To create a geofence via the dashboard, go to the Geofences page and click the New button. Optionally search for an address or place, then enter a description, optional tag, optional external ID, and optional metadata. Choose circle to create a circle geofence, polygon to create a polygon geofence, or isochrone to create an isochrone (time-based) geofence. Click Create to create the geofence.

CSV import

To create geofences via CSV import, go to the Geofences page and click the Import button. Then, select a CSV to upload.

The CSV must have 8 columns:

  • description (required): A display name for the geofence.
  • tag (optional, but recommended): A group for the geofence.
  • externalId (optional, but recommended): An external ID for the geofence that maps to your internal database.
  • type (required): The type of geofence geometry, either polygon, circle, or isochrone.
  • radius (required for type circle and isochrone): The radius in meters for type circle, a number between 50 and 10000. The travel duration in minutes and travel mode for type isochrone, a string in the format duration|mode (e.g., 15|car for 15 minutes driving). Ignored for type polygon.
  • coordinates (required): A JSON string representing a center in the format [longitude,latitude] for type circle. A JSON string representing a closed ring of between 4 and 10,000 coordinates in the format [[longitude0, latitude0],[longitude1,latitude1],[longitude2,latitude2],...,[longitude0,latitude0]] for type polygon. A JSON string representing a destination in the format [longitude,latitude] for type isochrone. Note that longitude comes before latitude, a GeoJSON convention.
  • enabled (optional): If true, the geofence will generate events. If false, the geofence will not generate events.
  • metadata (optional): A set of custom key-value pairs for the geofence. A JSON string representing a dictionary with up to 16 keys and values of type string, boolean, or number.

Headers for the columns should be omitted. In other words, the first row of the CSV should correspond to the first geofence.

If a geofence with the specified tag and external ID already exists, it will be updated. If not, it will be created.

For example, to import a circle geofence representing a store and a polygon geofence representing a neighborhood, you could use the following CSV:

"Store #123",store,123,circle,50,"[-73.975363,40.783826]",true,"{""parking"":false}"
"SoHo",neighborhood,soho,polygon,,"[[-73.996973,40.725660],[-73.998282,40.726294],[-74.001072,40.727660],[-74.002853,40.728603],[-74.003282,40.727936],[-74.003840,40.726310],[-74.004226,40.725351],[-74.004998,40.723530],[-74.005191,40.722993],[-74.005556,40.722115],[-74.004462,40.721334],[-74.001908,40.719496],[-73.999827,40.718016],[-73.997467,40.720863],[-73.997016,40.722212],[-73.996501,40.723497],[-73.995750,40.724278],[-73.995128,40.725107],[-73.996973,40.725660]]",true,
"15 minutes driving to Disney World","car-15",disney,isochrone,15|car,"[-81.563874,28.385233]",true,"{""type"":""resort""}"

Again, note that longitude comes before latitude, a GeoJSON convention.

API

Finally, you can create geofences programmatically via the API. You can create a geofence via POST /api/v1/geofences, or upsert a geofence based on tag and external ID via PUT /api/v1/geofences/:tag/:externalId.

For example, to upsert a geofence representing a store via the API:

curl https://api.radar.io/v1/geofences/store/123 \
    -H "Authorization: prj_live_sk_8ca5fdbe82ab7d49a1ca5c3f" \
    -X PUT \
    -d "description=Store #123" \
    -d "type=circle" \
    -d "coordinates=[-73.975363,40.783826]" \
    -d "radius=100"

Again, note that longitude comes before latitude, a GeoJSON convention.

Geofence sync

To automatically sync geofences nightly from a CSV, use geofence sync.

Confidence

All geofence events have confidence levels. Confidence levels range from 1 (low) to 3 (high).

When geofences do not overlap, confidence will always be high. When geofences do overlap, confidence will be medium or low for geofences with a radius of 200 meters or less, depending on the user's location, the user's location accuracy, and the geometry of the geofence.

You may decide to ignore some events based on confidence levels.

Stop detection

When Geofence Stop Detection is on, Radar can understand the difference between a user walking or driving through a geofence and stopping in a geofence, and will only generate a geofence entry event when a user stops in a geofence (i.e., when stopped is true based on tracking options or when foreground is true).

Turn on Geofence Stop Detection at a project level on the Settings page, under Geofences.