AR1/docs/analysis-tools/clustering.md

Clustering

Group features based on spatial proximity using clustering algorithms.

Overview

Clustering groups nearby features into clusters, identifying spatial patterns and groupings in point data.

Inputs

• Dataset: Point dataset
• Method: Clustering algorithm
• Parameters: Algorithm-specific parameters

Outputs

New dataset containing:

• Original features
• Cluster ID: Assigned cluster identifier
• Cluster Size: Number of features in cluster
• Original attributes

Algorithms

K-Means Clustering

Groups points into k clusters by minimizing within-cluster variance.

DBSCAN

Density-based clustering that identifies clusters of varying shapes.

Hierarchical Clustering

Builds cluster hierarchy using distance measures.

Example

{
  "dataset_id": 123,
  "method": "kmeans",
  "k": 5
}

Background Jobs

This analysis runs as a background job.

Use Cases

• Market segmentation
• Service area identification
• Pattern recognition
• Data exploration

Notes

• Algorithm selection depends on data characteristics
• Parameter tuning affects results
• Results may vary with different random seeds
• Consider spatial scale when interpreting clusters

Related Documentation

• Analysis API