Applications of DBSCAN · Real-World Use Cases

DBSCAN is one of the most cited algorithms in data mining (over 30,000 citations as of 2026). Its real-world applications span multiple domains.

Geographic Analysis

Crime hotspot detection: Chicago Police Department uses density-based clustering on geocoded incident reports to identify crime hotspots. DBSCAN’s ability to find arbitrarily-shaped clusters along transit corridors makes it more useful than k-means for this task.
Earthquake aftershock clustering: Seismology — aftershocks form spatial clusters that follow fault lines (linear, not circular). DBSCAN identifies these correctly.
GPS trajectory analysis: Identifying frequently visited locations from GPS traces. Stops form dense clusters; movement between locations is sparse.

Credit card fraud: Transactions outside normal spatiotemporal clusters are flagged.
Network intrusion detection: Normal traffic patterns form clusters; attacks appear in sparse regions.
Manufacturing quality control: Sensor readings from defective products fall outside the dense region of normal readings.

Behavioural clustering: Group customers by purchase patterns. DBSCAN separates loyal repeat-buyers (dense cluster) from one-off purchasers (noise) without forcing every customer into a segment.
Location-based segmentation: Retail site selection based on customer density.

Astronomy: Identifying star clusters, galaxy clusters, and stellar streams from survey data (Gaia, SDSS). Galaxy clusters are gravitationally bound — dense in 3D space — exactly what DBSCAN detects.
Bioinformatics: Clustering gene expression data to identify co-expressed gene modules. Protein structure clustering from molecular dynamics simulations.
Ecology: Identifying animal home ranges from GPS collar data.