point size 5
opacity 80%

What is UMAP?

UMAP (Uniform Manifold Approximation and Projection) finds a low-dimensional map of high-dimensional data that preserves its structure. It works by learning which points are neighbors in the original space, then arranging them so neighbors stay close in the map. Use the controls below to see how each parameter shapes the result.

Paper Docs
Method ?
Preprocessing
StandardScaler ?
Parameters
n_neighbors 15 ?
51015203050100
min_dist 0.1 ?
0.05.1.25.51
metric ?
Parameters
perplexity 30 ?
5153050100
metric ?
Components
show components ?
color_by ?
What is HDBSCAN?

HDBSCAN extends DBSCAN into a hierarchical clustering algorithm and extracts flat clusters based on their stability. It finds clusters of varying density without requiring you to specify the number of clusters in advance. Points that do not belong to any cluster are labeled noise (label −1).

Docs Paper
View ?
Color ?
Parameters
min_cluster_size ?
51015203050
min_samples ?
135101520
cluster_selection_method ?
cluster_selection_epsilon ?
0.1.25.5125
allow_single_cluster ?
cluster_on ?
Dataset
Python