Inspired by the excellent work the Information Interfaces Lab at Georgia Tech.
The little (green) dots are "dust" and represent individual data points and the larger circles are "magnets," each representing an attribute. The dust have many different attributes (each with separate, normalized values) and will gravitate towards magnets which they feel the strongest attraction to (highest value). This method allows for an intuitive method of visualizing relationships in multivariate data and finding unique points or interesting patterns.
We used Processing and just plain old Java to write a regular, vanilla desktop application. Call us boring or outdated but we like to stick to the tried and true. Our project might not be web-friendly and it might not be in node (the only real dev language) but it sure as hell works. It's functional, elegant, practical, beautiful and maybe if you really what to give it a spin we can email you a .jar.