Overview

Welcome to Multinet!

Multinet is a software suite consisting of a web API server, a web client application, a TypeScript client library, and a collection of analysis and visualization applications that are all designed to work with each other to host multivariate network data and enable discovery, analysis, and visualization for them.

Multinet is part of a larger NSF research project led by the Visualization Design Lab at the University of Utah, with Kitware Inc. spearheading software development efforts.

What is a Multivariate Network?

Networks are an increasingly common and useful way to model many of the phenomena in the natural, social, biological, and abstract worlds around us. Common examples include social networks connecting individuals through their acquaintanceships, neurons in the retina forming sensory-perceptual receiver circuits, computers connected to each other comprising both local networks and the Internet at large, and networks modeling the probabilistic transmission of disease through a community, but there are countless other examples, spread throughout many academic and industrial fields.

In general, networks are modeled as a collection of nodes, representing entities such as people or computers, connected together via links representing some type of relationship between a subset of pairs of nodes. Multivariate networks are networks that carry extra data (also known as attributes) on the nodes and edges. For instance, in a social network, the nodes (representing individual people) might each carry data such as name, age, hometown, etc. A network representing the retina could have scientific measurements or assessments, such as cell types for the nodes, or voltage measurements for the links, etc.

The Multinet research project’s main goal is to develop data processing, analysis, and visualization methods that take into account both the connectivity features of networks, and their multivariate data as well. Doing so will unlock a lot of the value of network datasets being collected by scientists, policy makers, and laypeople alike.

For an overview of the state of the art in visualization techniques for multivariate networks, see Nobre et al.

The Multinet Software Stack

The Multinet stack consists of the following components:

  • multinet-api, an open-source Django REST API server that provides the Multinet API, which we are developing as part of this project
  • multinet-client, an open-source web application, using VueJS and Vuetify to present an easy to use interface to the Multinet API, including access to stored data, and ways to launch intensive visualization and analysis applications
  • visualization applications, including:
    • MultiLink, an interactive node-link diagram that also displays node attributes in various forms
    • MultiMatrix, an interactive adjacency matrix that also supports attribute display
    • UpSet, a set visualization tool
  • ArangoDB, a third-party open-source graph database system that Multinet uses to store data and perform queries

You can see how the client works by going to https://multinet.app and giving it a try. To build your own development environment, either to host your own data locally, or to work on the codebases, see Developer Docs.