arcgis-spatial-analysis（arcgis 空间分析）

ArcGIS Spatial Analysis

Use this skill for spatial analysis objects (3D analyses), feature queries with statistics, and feature reduction (clustering/binning).

For geometry operators (buffer, union, intersect, clip, offset, etc.), see arcgis-geometry-operations. For routing, geocoding, and geoprocessing REST services, see arcgis-rest-services.

Import Patterns

Direct ESM Imports

import ViewshedAnalysis from "@arcgis/core/analysis/ViewshedAnalysis.js";
import Viewshed from "@arcgis/core/analysis/Viewshed.js";
import LineOfSightAnalysis from "@arcgis/core/analysis/LineOfSightAnalysis.js";
import ElevationProfileAnalysis from "@arcgis/core/analysis/ElevationProfileAnalysis.js";
import SliceAnalysis from "@arcgis/core/analysis/SliceAnalysis.js";
import ShadowCastAnalysis from "@arcgis/core/analysis/ShadowCastAnalysis.js";

Dynamic Imports (CDN)

const [ViewshedAnalysis, Viewshed] = await $arcgis.import([
  "@arcgis/core/analysis/ViewshedAnalysis.js",
  "@arcgis/core/analysis/Viewshed.js",
]);

Analysis Objects (3D)

Analysis objects are added to a SceneView's analyses collection. All analysis types require SceneView (not MapView).

ElevationProfileAnalysis

import ElevationProfileAnalysis from "@arcgis/core/analysis/ElevationProfileAnalysis.js";
import * as reactiveUtils from "@arcgis/core/core/reactiveUtils.js";

const analysis = new ElevationProfileAnalysis({
  profiles: [
    { type: "ground", color: "brown" },
    { type: "input", color: "blue" },
  ],
  displayUnits: {
    distance: "meters",
    elevation: "meters",
  },
});

view.analyses.add(analysis);
analysis.geometry = polyline;

// Get analysis view for results
const analysisView = await view.whenAnalysisView(analysis);

reactiveUtils.watch(
  () => analysisView.progress,
  (progress) => {
    if (progress === 1) {
      console.log("Results:", analysisView.results);
    }
  },
);

Interactive Analysis Placement

Analysis views support interactive placement via the place() method:

import * as promiseUtils from "@arcgis/core/core/promiseUtils.js";

const analysisView = await view.whenAnalysisView(analysis);
const abortController = new AbortController();

try {
  await analysisView.place({ signal: abortController.signal });
} catch (error) {
  if (!promiseUtils.isAbortError(error)) throw error;
}

// To cancel placement
abortController.abort();

ViewshedAnalysis

Uses a viewsheds collection of Viewshed objects (not direct observer property).

import ViewshedAnalysis from "@arcgis/core/analysis/ViewshedAnalysis.js";
import Viewshed from "@arcgis/core/analysis/Viewshed.js";

const viewshed = new Viewshed({
  observer: {
    type: "point",
    x: -122.4,
    y: 37.8,
    z: 100,
    spatialReference: { wkid: 4326 },
  },
  farDistance: 1000,
  heading: 45,
  tilt: 90,
  horizontalFieldOfView: 120,
  verticalFieldOfView: 90,
});

const analysis = new ViewshedAnalysis({
  viewsheds: [viewshed],
});

view.analyses.add(analysis);

// Add more viewsheds dynamically
analysis.viewsheds.add(
  new Viewshed({
    observer: {
      type: "point",
      x: -122.41,
      y: 37.79,
      z: 80,
      spatialReference: { wkid: 4326 },
    },
    farDistance: 500,
    heading: 0,
    tilt: 90,
    horizontalFieldOfView: 360,
    verticalFieldOfView: 90,
  }),
);

// Clear all viewsheds
analysis.clear();

LineOfSightAnalysis

Uses LineOfSightAnalysisObserver and LineOfSightAnalysisTarget wrapper objects.

import LineOfSightAnalysis from "@arcgis/core/analysis/LineOfSightAnalysis.js";
import LineOfSightAnalysisObserver from "@arcgis/core/analysis/LineOfSightAnalysisObserver.js";
import LineOfSightAnalysisTarget from "@arcgis/core/analysis/LineOfSightAnalysisTarget.js";

const analysis = new LineOfSightAnalysis({
  observer: new LineOfSightAnalysisObserver({
    position: {
      type: "point",
      x: -122.4,
      y: 37.8,
      z: 100,
      spatialReference: { wkid: 4326 },
    },
  }),
  targets: [
    new LineOfSightAnalysisTarget({
      position: {
        type: "point",
        x: -122.41,
        y: 37.81,
        z: 50,
        spatialReference: { wkid: 4326 },
      },
    }),
    new LineOfSightAnalysisTarget({
      position: {
        type: "point",
        x: -122.42,
        y: 37.79,
        z: 75,
        spatialReference: { wkid: 4326 },
      },
    }),
  ],
});

view.analyses.add(analysis);

ShadowCastAnalysis

Uses date, startTimeOfDay, endTimeOfDay, and mode properties.

import ShadowCastAnalysis from "@arcgis/core/analysis/ShadowCastAnalysis.js";

const analysis = new ShadowCastAnalysis({
  date: new Date("2025-06-21"),
  startTimeOfDay: 21600000, // 6:00 AM (milliseconds from midnight)
  endTimeOfDay: 57600000, // 4:00 PM
  mode: "total-duration", // "total-duration", "discrete", "min-duration"
  utcOffset: -7, // UTC offset in hours
});

view.analyses.add(analysis);

SliceAnalysis

Uses a shape property of type SlicePlane (not plane).

import SliceAnalysis from "@arcgis/core/analysis/SliceAnalysis.js";
import SlicePlane from "@arcgis/core/analysis/SlicePlane.js";

const analysis = new SliceAnalysis({
  shape: new SlicePlane({
    position: {
      type: "point",
      x: -122.4,
      y: 37.8,
      z: 50,
      spatialReference: { wkid: 4326 },
    },
    heading: 0,
    tilt: 0,
    width: 200,
    height: 200,
  }),
  tiltEnabled: false,
  excludeGroundSurface: false,
});

view.analyses.add(analysis);

AreaMeasurementAnalysis

import AreaMeasurementAnalysis from "@arcgis/core/analysis/AreaMeasurementAnalysis.js";

const analysis = new AreaMeasurementAnalysis();
view.analyses.add(analysis);

// Set geometry after user draws or programmatically
analysis.geometry = polygon;

// Watch for results
const analysisView = await view.whenAnalysisView(analysis);
reactiveUtils.watch(
  () => analysisView.result,
  (result) => {
    if (result) {
      console.log("Area:", result.area?.value, result.area?.unit);
      console.log(
        "Perimeter:",
        result.perimeterLength?.value,
        result.perimeterLength?.unit,
      );
    }
  },
  { initial: true },
);

DistanceMeasurementAnalysis

Uses a geometry property (Polyline), not startPoint/endPoint.

import DistanceMeasurementAnalysis from "@arcgis/core/analysis/DistanceMeasurementAnalysis.js";

const analysis = new DistanceMeasurementAnalysis({
  geometry: {
    type: "polyline",
    paths: [
      [
        [-122.4, 37.8, 0],
        [-122.41, 37.81, 0],
      ],
    ],
    spatialReference: { wkid: 4326 },
  },
});

view.analyses.add(analysis);

VolumeMeasurementAnalysis

import VolumeMeasurementAnalysis from "@arcgis/core/analysis/VolumeMeasurementAnalysis.js";

const analysis = new VolumeMeasurementAnalysis();
view.analyses.add(analysis);

analysis.geometry = polygon;

// Watch for results
const analysisView = await view.whenAnalysisView(analysis);
reactiveUtils.watch(
  () => analysisView.result,
  (result) => {
    if (result) {
      console.log(
        "Cut volume:",
        result.cutVolume?.value,
        result.cutVolume?.unit,
      );
      console.log(
        "Fill volume:",
        result.fillVolume?.value,
        result.fillVolume?.unit,
      );
    }
  },
  { initial: true },
);

DimensionAnalysis

import DimensionAnalysis from "@arcgis/core/analysis/DimensionAnalysis.js";
import LengthDimension from "@arcgis/core/analysis/LengthDimension.js";

const dimension = new LengthDimension({
  startPoint: {
    type: "point",
    x: -122.4,
    y: 37.8,
    z: 0,
    spatialReference: { wkid: 4326 },
  },
  endPoint: {
    type: "point",
    x: -122.41,
    y: 37.81,
    z: 0,
    spatialReference: { wkid: 4326 },
  },
});

const analysis = new DimensionAnalysis({
  dimensions: [dimension],
});

view.analyses.add(analysis);

Analysis Types Summary

Feature Queries with Statistics

Basic Statistics

const query = featureLayer.createQuery();
query.outStatistics = [
  {
    statisticType: "sum",
    onStatisticField: "population",
    outStatisticFieldName: "totalPop",
  },
  {
    statisticType: "avg",
    onStatisticField: "population",
    outStatisticFieldName: "avgPop",
  },
  {
    statisticType: "max",
    onStatisticField: "population",
    outStatisticFieldName: "maxPop",
  },
  {
    statisticType: "min",
    onStatisticField: "population",
    outStatisticFieldName: "minPop",
  },
  {
    statisticType: "count",
    onStatisticField: "population",
    outStatisticFieldName: "count",
  },
  {
    statisticType: "stddev",
    onStatisticField: "population",
    outStatisticFieldName: "stdDev",
  },
];

const result = await featureLayer.queryFeatures(query);
console.log(result.features[0].attributes);

Group By Statistics

const query = featureLayer.createQuery();
query.groupByFieldsForStatistics = ["state"];
query.outStatistics = [
  {
    statisticType: "sum",
    onStatisticField: "population",
    outStatisticFieldName: "totalPop",
  },
];
query.orderByFields = ["totalPop DESC"];

const result = await featureLayer.queryFeatures(query);
// Returns one feature per state with total population

Spatial Statistics

const query = featureLayer.createQuery();
query.geometry = view.extent;
query.spatialRelationship = "intersects";
query.outStatistics = [
  {
    statisticType: "count",
    onStatisticField: "ObjectID",
    outStatisticFieldName: "featureCount",
  },
];

const result = await featureLayer.queryFeatures(query);
console.log("Features in view:", result.features[0].attributes.featureCount);

Feature Reduction

Clustering

const clusterConfig = {
  type: "cluster",
  clusterRadius: "100px",
  clusterMinSize: "24px",
  clusterMaxSize: "60px",
  popupTemplate: {
    title: "Cluster summary",
    content: "This cluster represents {cluster_count} features.",
    fieldInfos: [
      {
        fieldName: "cluster_count",
        format: { digitSeparator: true, places: 0 },
      },
    ],
  },
  labelingInfo: [
    {
      deconflictionStrategy: "none",
      labelExpressionInfo: {
        expression: "Text($feature.cluster_count, '#,###')",
      },
      symbol: {
        type: "text",
        color: "white",
        font: { size: "12px", weight: "bold" },
      },
      labelPlacement: "center-center",
    },
  ],
};

featureLayer.featureReduction = clusterConfig;

// Toggle clustering
featureLayer.featureReduction = null; // Disable
featureLayer.featureReduction = clusterConfig; // Enable

Cluster with Aggregated Fields

const clusterConfig = {
  type: "cluster",
  clusterRadius: "100px",
  fields: [
    {
      name: "avg_magnitude",
      statisticType: "avg",
      onStatisticField: "magnitude",
    },
    {
      name: "total_count",
      statisticType: "count",
      onStatisticField: "ObjectID",
    },
  ],
  renderer: {
    type: "simple",
    symbol: {
      type: "simple-marker",
      style: "circle",
      color: "#69dcff",
    },
    visualVariables: [
      {
        type: "size",
        field: "total_count",
        stops: [
          { value: 1, size: 8 },
          { value: 100, size: 40 },
        ],
      },
    ],
  },
};

Binning

const binConfig = {
  type: "binning",
  fixedBinLevel: 4,
  renderer: {
    type: "simple",
    symbol: {
      type: "simple-fill",
      outline: { color: "white", width: 0.5 },
    },
    visualVariables: [
      {
        type: "color",
        field: "aggregateCount",
        stops: [
          { value: 1, color: "#feebe2" },
          { value: 50, color: "#fbb4b9" },
          { value: 100, color: "#f768a1" },
          { value: 500, color: "#c51b8a" },
          { value: 1000, color: "#7a0177" },
        ],
      },
    ],
  },
  popupTemplate: {
    title: "Bin",
    content: "{aggregateCount} features in this bin",
  },
};

featureLayer.featureReduction = binConfig;

Common Pitfalls

1. Analysis objects are SceneView-only: All analysis types (viewshed, line of sight, shadow cast, etc.) require 3D SceneView. They do not work in MapView.
2. ViewshedAnalysis uses a viewsheds collection: Do not set observer directly on ViewshedAnalysis. Create Viewshed objects and add them to the viewsheds collection: // Anti-pattern new ViewshedAnalysis({observer: point, farDistance: 1000}); // Correct new ViewshedAnalysis({viewsheds: [new Viewshed({observer: point, farDistance: 1000})],});
3. Cluster field names: Use cluster_count (not clusterCount) to access the feature count in cluster popups and labels.
4. Statistics queries return features: Statistics queries return features with calculated attributes, not raw numbers. Access results via result.features[0].attributes.
5. Analysis progress: Check analysisView.progress === 1 or watch analysisView.result to know when analysis results are ready. Don't read results before they're computed.
6. SliceAnalysis uses shape not plane: The property is shape of type SlicePlane, not plane: // Anti-pattern new SliceAnalysis({plane: slicePlane}); // Correct new SliceAnalysis({shape: slicePlane});

Reference Samples

• analysis-viewshed - Interactive viewshed analysis in 3D
• analysis-objects - Multiple analysis types in a single scene
• analysis-elevation-profile - Elevation profile along a line
• analysis-shadow-cast - Shadow cast duration visualization
• analysis-area-measurement - Area measurement on 3D surfaces
• analysis-volume-measurement - Cut/fill volume measurement
• line-of-sight - Line of sight visibility analysis
• featurereduction-cluster - Feature clustering
• featurereduction-binning - Feature binning for aggregation
• featurereduction-cluster-aggregate-fields - Clusters with aggregated statistics

Related Skills

• See arcgis-geometry-operations for client-side geometry operations (buffer, union, intersect, etc.).
• See arcgis-rest-services for server-side REST services (routing, geocoding, geoprocessing).
• See arcgis-coordinates-projection for coordinate projection and formatting.
• See arcgis-layers for FeatureLayer queries and layer configuration.