Clustering

Global power plants clustered and categorized by fuel type. Clustering summarizes the layer's renderer so you can see the spatial density of features at a glance.

What is clustering?

Clustering is a method of reducing points in a layer by grouping them into clusters based on their spatial proximity to one another. Typically, clusters are proportionally sized based on the number of features within each cluster.

This is an effective way to show areas where many points stack on top of one another.

Clustering allows you to effectively visualize where points stack on top of another or are in very close proximity to each other. Use the swipe widget above to compare an unclustered layer of power plants with a clustered version.

Why is clustering useful?

Large point layers can be deceptive. What appears to be just a few points can in reality be several thousand. Clustering allows you to visually represent large numbers of points in relatively small areas.

For example, the following map shows the locations of thousands of power plants. In the image below, regions A and B both have a high density of points, making them impossible to compare.

Region A and region B both have a high density of points. It is impossible to tell how many points overlap in each area.

However, when clustering is enabled, the user can now clearly see that region B has nearly twice as many points as region A.

Clustering allows the user to easily compare the density of overlapping features at a glance.

How clustering works

Clustering is configured on the featureReduction property of the layer. You can enable clustering with minimal code by setting the featureReduction type to cluster.

Use dark colors for code blocksCopy
layer.featureReduction = {
  type: "cluster"
};

The featureReduction property gives you control over many other cluster properties. The clusterRadius defines each cluster's area of influence for including features. You may also define popupTemplates and labels for clusters to summarize the features included in the cluster.

Examples

Basic clustering

The following example demonstrates how to enable clustering on a point layer and configure labels and a popup for displaying the cluster count.

The aggregate fields used by clusters are generated once clustering is enabled on the layer. By default, all clustered layers have a cluster_count aggregate field. This can be used in the labels and the popup for each cluster. Other fields used in the layer's renderer may be accessible for display in the popup. You can learn more about how to use these in the FeatureReductionCluster.popupTemplate documentation.

See the Related samples and resources below for more examples of how to summarize data within a cluster's popup.

Global power plants clustered by count.

ArcGIS JS API
Expand
Use dark colors for code blocksCopy
52
52
52
52
52
52
52
52
52
52
52
52
52
52
52
52
52
52
52
52
52
52
52
52
52
52
52
52
52
52
52
52
52
52
52
52
52
52
52
52
52
52
52
52
52
52
52
52
52
52
52
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
86
86
86
86
86
86
86
86
86
86
86
86
86
86
86
86
86
86
86
86
86
86
86
86
86
86
86
86
86
86
86
86
86
86
86
86
86
86
86
86
86
86
86
86
        clusteredLayer.featureReduction = {
          type: "cluster",
          clusterMinSize: 16.5,
          // defines the label within each cluster
          labelingInfo: [
            {
              deconflictionStrategy: "none",
              labelExpressionInfo: {
                expression: "Text($feature.cluster_count, '#,###')"
              },
              symbol: {
                type: "text",
                color: "white",
                font: {
                  family: "Noto Sans",
                  size: "12px"
                }
              },
              labelPlacement: "center-center"
            }
          ],
          // information to display when the user clicks a cluster
          popupTemplate: {
            title: "Cluster Summary",
            content: "This cluster represents <b>{cluster_count}</b> features.",
            fieldInfos: [{
              fieldName: "cluster_count",
              format: {
                places: 0,
                digitSeparator: true
              }
            }]
          }
        };
Expand

Suggested cluster defaults

By default, the cluster symbol always summarizes the features in the cluster. When a layer has a UniqueValueRenderer, the symbol of each cluster represents the predominant value of features in the cluster. When a layer has any visual variables applied to it, the average of each variable in the cluster is applied to the cluster symbol. The fields describing the predominant type and average of numeric fields can be referenced in the cluster popup and label.

This example uses smart mapping methods to demonstrate how to generate the suggested cluster configuration specific to the layer's renderer.

Global power plants clustered and categorized by fuel type. Clustering summarizes the layer renderer so you can see a summary of the features contained by the cluster at a glance.

ArcGIS JS API
Expand
Use dark colors for code blocksCopy
102
102
102
102
102
102
102
102
102
102
102
102
102
102
102
102
102
102
102
102
102
102
102
102
102
102
102
102
102
102
102
102
102
102
102
102
102
102
102
102
102
102
102
102
102
102
102
102
102
102
102
102
102
102
102
102
102
102
102
102
102
102
102
102
102
102
102
102
102
102
102
102
102
102
102
102
102
102
102
102
102
102
102
102
102
102
102
102
102
102
102
102
102
102
102
102
102
102
102
102
102
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
122
122
122
122
122
122
122
122
122
122
122
          // generates default popupTemplate
          const popupTemplate = await clusterPopupCreator
            .getTemplates({ layer })
            .then( (popupTemplateResponse) => popupTemplateResponse.primaryTemplate.value );

          // generates default labelingInfo
          const { labelingInfo, clusterMinSize } = await clusterLabelCreator
            .getLabelSchemes({
              layer,
              view
            })
            .then((labelSchemes) => labelSchemes.primaryScheme);

          // Set this object on layer.featureReduction
          return {
            type: "cluster",
            popupTemplate,
            labelingInfo,
            clusterMinSize
          };
Expand

Clusters as pie charts

By default, when a layer has a UniqueValueRenderer or ClassBreaksRenderer, the symbol of each cluster represents the predominant category within the cluster.

You may prefer to visualize clusters of type-based renderers as pie charts rather than the predominant category. To do this, you can call the pieChart.createRendererForClustering method, which will create a pie chart renderer based on the categories defined in a UniqueValueRenderer or a ClassBreaksRenderer.

The fields and renderer returned from this method should be set directly on the FeatureReductionCluster instance on the layer.

Use dark colors for code blocksCopy
const { renderer, fields } = await pieChartRendererCreator.createRendererForClustering({
  layer,
  shape: "donut"
});

layer.featureReduction = {
  type: "cluster",
  fields,
  renderer
}

311 incident reports categorized by fuel type. Clustering summarizes the layer renderer as a pie chart for layers with a UniqueValueRenderer or ClassBreaksRenderer.

Intro to clustering

Override cluster symbol

Clusters as pie charts

Cluster size based on the sum of a field

Clustering with aggregate fields

Clustering - generate suggested configuration

Clustering - filter popup features

This sample demonstrates how to filter clustered features within a cluster's popup.

Clustering - query clusters

Clustering - advanced configuration

Point clustering with visual variables

FeatureReductionCluster

Read the Core API Reference for more information.

API support

The following table describes the geometry and view types that are suited well for each visualization technique.

	2D	Lines	Polygons	Client-side	Server-side
Clustering
Binning
Heatmap
Opacity
Bloom
Aggregation
Thinning	¹	¹	¹	²	³
Visible scale range

Full supportPartial supportNo support

1. Feature reduction selection not supported
2. Only by feature reduction selection
3. Only by scale-driven filter