2 Comments

Summary:

The tool, which forms part of Recommind’s cloud-based Axcelerate On-Demand package, aims to give non-technical users a faster and more informative e-discovery process.

lawyer

If you’re a lawyer going through the discovery process, or you’re just into playing NSA with your employees’ communications,  this one’s for you: Recommind, the e-discovery and enterprise search firm, just launched a new tool called Hypergraph for discovering and visualizing patterns in unstructured data.

Hypergraph, which is now part of Recommind’s Axcelerate On-Demand SaaS package, is a tool for finding and highlighting hidden connections between people, documents, messages and other elements. As with so many data visualization tools, the idea is to make it easier for even non-technical people to look at a mass of unstructured data and spot the links they need to find – such as, for example, a naughty employee’s telltale communications with the wrong people.

Here’s the sort of visualization we’re talking about:

Recommind Hypergraph

Of course, Recommind does have competition in this space, notably from ZyLab, EMC and IBM. However, the company is pushing the ability of Hypergraph to parse and provide insights into big data, something it does through machine learning and graph analysis techniques. That in turn makes for more informative visualizations.

As Recommind CEO Bob Tennant pitches it:

“Unstructured data represents the great majority of the information universe, and it’s by far the hardest to analyse. Data visualization in combination with data mining can produce insight that would be hard to achieve any other way. To make a significant impact, visualisation has to be massively scalable, highly automated and simple to use. Hypergraph is the first solution that has it all.”

  1. Very impressive. Data visualization is nothing new, but applying predictive analytics to it is potenially huge. I’m very interested in seeing where this goes.

    Share
  2. These guys have been THE name in predictive coding for so long, I’m actually rather surprised it took them this long to figure out that they can apply the same principles to analyze Big Data. I can see something like this working great against a Hadoop cluster to finally solve the “correlation” vs. “causality” problem we all experience.

    Share

Comments have been disabled for this post