DBSCAN (Density-Based Spatial Clustering) › Visualizing DBSCAN

Michael BrenndoerferUpdated December 18, 20255 min read
Machine Learning from ScratchAll chapters
Part 7 of 9Unsupervised Learning
Chapter 55DBSCAN (Density-Based Spatial Clustering) › Visualizing DBSCAN
Reading section 3 of 8 · Visualizing DBSCAN · 5 minSection 3 of 8 · 5 min

Book contents

Machine Learning from Scratch9 parts · 66 chaptersView the whole book →
Active chapter 55 of 66

Section 3 of 8 · 5 min read

DBSCAN (Density-Based Spatial Clustering) › Visualizing DBSCANOne section · Sections in this chapter3 / 8 ↓
Sections in this chapter8
Reading tools Definitions & inline links

Choose your expertise level to adjust how many terms are explained. Beginners see more tooltips, experts see fewer to maintain reading flow. Hover over underlined terms for instant definitions.

Article links

Make inline references clickable

Let's create visualizations that demonstrate how DBSCAN works with different types of data and parameter settings. We'll show the algorithm's ability to find clusters of arbitrary shapes and handle noise effectively.

Out[5]:
Visualization
Two crescent-shaped moon clusters with scattered noise points, showing non-spherical cluster geometry.
Original two moons dataset showing crescent-shaped clusters with noise. This dataset demonstrates DBSCAN's ability to identify non-spherical clusters that would be challenging for centroid-based methods like k-means. The two moon shapes are clearly separated but have complex curved boundaries that require density-based clustering to identify correctly.
Nested concentric circles forming two separate circular clusters at different radii.
Original concentric circles dataset with two circular clusters at different radii. This dataset tests DBSCAN's ability to handle nested cluster structures where one cluster completely surrounds another. The algorithm must distinguish between the inner and outer circles while avoiding merging them into a single cluster.
Four blob clusters with varying densities and sizes, testing DBSCAN's robustness to density variations.
Original blobs dataset with varying cluster densities. This dataset contains four clusters with different standard deviations (1.0, 2.5, 0.5, 1.5), creating clusters of varying sizes and densities. This tests DBSCAN's sensitivity to density variations and its ability to handle clusters with different characteristics in the same dataset.
Out[6]:
Visualization
DBSCAN successfully separates two moon-shaped clusters with noise points correctly identified as outliers.
DBSCAN clustering results on the two moons dataset. The algorithm successfully identifies the two crescent-shaped clusters (shown in different colors) while correctly classifying noise points as outliers (black 'x' markers). This demonstrates DBSCAN's strength in handling non-spherical clusters and automatic noise detection.
DBSCAN maintains separation between inner and outer circular clusters despite nested structure.
DBSCAN clustering results on the concentric circles dataset. The algorithm correctly separates the inner and outer circular clusters despite their nested structure. This shows DBSCAN's ability to handle complex spatial relationships and maintain cluster boundaries even when clusters are not linearly separable.
DBSCAN identifies all four clusters despite significant density and size variations.
DBSCAN clustering results on the varying density blobs dataset. The algorithm identifies all four clusters despite their different densities and sizes. This demonstrates DBSCAN's robustness to density variations and its ability to find clusters with different characteristics in the same dataset.

Here you can see DBSCAN's key strengths in action: it can identify clusters of arbitrary shapes (moons, circles) and handle varying densities while automatically detecting noise points (shown in black with 'x' markers). The algorithm successfully separates the two moon-shaped clusters, identifies the concentric circular patterns, and finds clusters with different densities in the blob dataset.