Dataset

Upload CSV or Excel (.xlsx) file

Try with a sample dataset

Pick one to explore instantly — no upload needed

Outlier Detection

GLOSH + probability composite scoring

Min Cluster Size

Hierarchical Density

Primary HDBSCAN parameter – intuitive cluster size threshold

Minimum cluster size5

Guidelines:

• Small (2-5): Finds small, fine-grained clusters

• Medium (10-25): Balanced for most datasets

• Large (50+): Only finds very significant clusters

Parameter Sweep

Min Samples (Density)

Controls density estimate – affects noise sensitivity

Auto (= min_cluster_size)

Cluster Selection

How to extract flat clustering from hierarchy

Selection method

EOM: Finds clusters of varying sizes. Best for heterogeneous data where cluster sizes differ. Tends to find larger, more stable clusters. Use for general purpose clustering.

Selection epsilon (ε)0.00

Merge clusters below this distance. 0 = stability-based selection (default). Higher = fewer, larger clusters.

Distance Metric

How to measure point similarity

Euclidean works well for scaled numeric data. Use Cosine for high-dimensional or text-based features.

Preprocessing

Recommended feature transformations

Enable preprocessing

Advanced settings

Projection

Choose embedding for 2D visualization

Projection method

PCA is deterministic and highlights global variance. UMAP captures complex manifolds for exploratory analysis.

Parameter Guide

Click to expand

Performance Metrics

Min Cluster Size

—

Primary HDBSCAN param

Clusters

—

Points: 0

DBCV Score

—

Density-based validity

High-Confidence

—

Strong assignments

Noise Points

—

Outliers

Selection

—

Excess of Mass

Interactive Visualization

Cluster Visualization

Interactive scatter plot with cluster centroids

Run training to render the scatter plot

Advanced Analysis

Deep Dive Analysis

Advanced metrics and interpretation tools

Recent Jobs

Job History

Recent jobs — auto-pruned to keep 10 per type

No jobs recorded yet. Run a clustering, classification, or pipeline job to see results here.

Loading…

Dataset

Upload CSV or Excel (.xlsx) file

Try with a sample dataset

Pick one to explore instantly — no upload needed

Outlier Detection

GLOSH + probability composite scoring

Min Cluster Size

Hierarchical Density

Primary HDBSCAN parameter – intuitive cluster size threshold

Minimum cluster size5

Guidelines:

• Small (2-5): Finds small, fine-grained clusters

• Medium (10-25): Balanced for most datasets

• Large (50+): Only finds very significant clusters

Parameter Sweep

Min Samples (Density)

Controls density estimate – affects noise sensitivity

Auto (= min_cluster_size)

Cluster Selection

How to extract flat clustering from hierarchy

Selection method

EOM: Finds clusters of varying sizes. Best for heterogeneous data where cluster sizes differ. Tends to find larger, more stable clusters. Use for general purpose clustering.

Selection epsilon (ε)0.00

Merge clusters below this distance. 0 = stability-based selection (default). Higher = fewer, larger clusters.

Distance Metric

How to measure point similarity

Euclidean works well for scaled numeric data. Use Cosine for high-dimensional or text-based features.

Preprocessing

Recommended feature transformations

Enable preprocessing

Advanced settings

Projection

Choose embedding for 2D visualization

Projection method

PCA is deterministic and highlights global variance. UMAP captures complex manifolds for exploratory analysis.

Parameter Guide

Click to expand

Performance Metrics

Min Cluster Size

—

Primary HDBSCAN param

Clusters

—

Points: 0

DBCV Score

—

Density-based validity

High-Confidence

—

Strong assignments

Noise Points

—

Outliers

Selection

—

Excess of Mass

Interactive Visualization

Cluster Visualization

Interactive scatter plot with cluster centroids

Run training to render the scatter plot

Advanced Analysis

Deep Dive Analysis

Advanced metrics and interpretation tools

Recent Jobs

Job History

Recent jobs — auto-pruned to keep 10 per type

No jobs recorded yet. Run a clustering, classification, or pipeline job to see results here.

HDBSCAN Clustering — Interactive Analysis