DBSCAN-visualiser

Demo

https://sheltered-retreat-53924.herokuapp.com

Description

Shows the DBSCAN clustering algorithm in action

Backend: Spring Boot Web

Frontend: p5js

Example url: http://localhost:8080/?eps=0.05&pts=5&gap=0.3&innerSize=50&outerSize=1300

The clustering algorithm is DBSCAN. I use the apache commons-math3 library to cluster the data.

You can set the eps and pts attributes of the DBSCAN algorithm. Please keep in mind that the data collection is generated at the beggining of the application and it does not change in time. !UPDATE! You can regenerate the data whenever you want with the 'ReGenerate' button on the UI.

When you modify the attributes of the clustering algorithm, in the background the backend uses the same data collection and run the clustering with the new attrbitues.

Screenshots

2016.11.06

Clustering

Drawer Preview before Uploading - You can edit the data!

2016.11.01

2016.10.31

TODO

• Backend

• API to regenerate the data collection
• API to set the attributes of the data collection generation, like gap, inner radius, outer radius
• API to consume custom data points
• Add option to use Perlin noise or other random generator
• Add more example data collection
  • Frontend
• Add 'Regenerate' button which regenerates the data collection
• Add inputs to set the data generation's attributes
• Add a selector for more example data collection
• URL should contains the settings (parameters of clustering, example data) and use them
• Add textarea for the custom data points
• Add option to use Perlin noise or other random generator
• Create subpage - Cluster Drawer - Create data points with mouse
• Add graph statistics - density, min and max distance, etc.
• Visualise the data if that has more than 2 dimensions