K-Means as a Service

This service analyzes data to determine the best number of clusters for it, using K-Means algorithm with Mahalanois distance and Bayesian Information Criterion.

New Analysis

Upload your data file here to start a new analysis:

Number of times the analysis is run to create confidence intervals. Min: 1. Max: 100.
Maximum number of clusters to fit to the data. Min: 1. Max: 15.
Number of times to initialize each fit. Min: 1. Max: 100.
Covariance matrices to use for the analysis.
Scale the data so that each dimension has zero mean and unit standard deviation. This option is recommended.
Only ".csv" files smaller than 256MB in size are allowed. The files must have a header with column names.
Columns from the file to be used for clustering. "Longitude" and "Latitude" are excluded by default, but can be added above.

Check Status

Check the status of a previously submitted job here:

View Report

View the report for a previously submitted job here:


Demonstration of the usefullness of this analysis on synthetic datasets:

Skewed 1

View Report »

Skewed 2


Skewed 3

View Report »

Skewed 4

View Report »

Skewed 5


Overlap 1

sewing board

Overlap 2

(518) 450-7418