What is a cluster analysis?
A cluster analysis is the process of finding clusters of point features within surrounding noise at a specified distance. To execute the analysis, use the spatial analysis service and the Find
operation.
Real-world examples of this analysis include the following:
- Where are there clusters of traffic incidents?
- Where are there clusters of insect infestations?
- Where are there clusters of trees?
How to perform a point cluster analysis
- Review the parameters for the
Find
operation.Point Clusters - Send a request to get the spatial analysis service URL.
- Execute a job request with the following URL and parameters:
- URL:
https
:// <YOUR _ANALYSIS _SERVICE >/arcgis/rest/services/tasks/ GP Server/ Find Point Clusters/submit Job - Parameters:
analysis
: Your dataset as a hosted feature layer or feature collection.Layer min
: Define how many points constitute a cluster.Features Cluster output
: A string representing the name of the hosted feature layer to return with the results.Name
- URL:
- Check the status.
- Get the output layer results.
To see examples using ArcGIS API for Python, ArcGIS REST JS, and the ArcGIS REST API, go to Examples below.
URL request
http://<YOUR_ANALYSIS_SERVICE>/arcgis/rest/services/tasks/GPServer/FindPointClusters/submitJob?<parameters>
Required parameters
Name | Description | Examples |
---|---|---|
f | The format of the data returned. | f=json f=pjson |
token | An OAuth 2.0 access token. | token= |
analysis | The point data. | {url |
min | Define the minimum number of features to be considered a cluster. | 5 |
Key parameters
Name | Description | Examples |
---|---|---|
search | A double value to use as the maximum distance to search for neighboring features. | 200 |
search | Must be provided if search is specified. | Miles , Feet , Meters , Kilometers |
output | A string representing the name of the hosted feature layer to return with the results. NOTE: If you do not include this parameter, the results are returned as a feature collection (JSON). | {"service |
context | A bounding box or output spatial reference for the analysis. | "extent" |
Example
Find clusters of traffic crashes
This example uses the Find
operation to find where there are areas (clusters) of traffic crashes. The analysis
value is the Traffic crashes hosted feature layer. The min
is defined as 500
. The results show four distinct colors that represent the four distinct clusters found in the analysis.
APIs
analysisLayer = "https://services3.arcgis.com/GVgbJbqm8hXASVYi/arcgis/rest/services/Traffic_Crashes/FeatureServer/0"
results = find_point_clusters(
analysis_layer=analysisLayer,
min_features_cluster=500,
#Ouputs results as a hosted feature service.
output_name="Find point clusters results",
)
result_features = results.layers[0].query()
print(
f"The find point clusters layer has {len(result_features.features)} new records"
)
Service requests
Request
POST arcgis.com/sharing/rest/portals/self HTTP/1.1
Content-Type: application/x-www-form-urlencoded
&f=json
&token=<ACCESS_TOKEN>
Response (JSON)
{
"helperServices": {
// Other parameters
"analysis": {
"url": "https://<YOUR_ANALYSIS_SERVICE>/arcgis/rest/services/tasks/GPServer"
},
"geoenrichment": {
"url": "https://geoenrich.arcgis.com/arcgis/rest/services/World/GeoenrichmentServer"
}
}
}