Easy Workflow Wins With Text Analytics



by Constantine Pappas on December 18, 2014

Analytics & Assisted Review , Litigation Support , Product Spotlight

We had the pleasure and privilege to offer many outstanding, well-received sessions at Relativity Fest this year, but the one that I had a special affinity for was the Quick Wins with Analytics breakout. Why? Simply put, there are tools out there that many folks are unaware of that could make their jobs—and lives—much easier.

As technology like Relativity Assisted Review and email threading become more popular in the industry for their efficacy, it seems only natural that some of their little sibling text analytics features would feel a bit neglected.

To tackle that gap, we want to delve into the corners of the Relativity Analytics toolbox and let our users know just what they can accomplish with the product’s other features. When I polled Relativity Fest attendees about why their teams and customers are resistant to using analytics, the two primary reasons provided are cost, and a fear of the unknown or lack of understanding.

These reasons feed into each other. After all, it can be a challenge to recognize something’s value if you lack a basic understanding of its function. That’s where the following tools come in. They have simple, intuitive workflows that yield immediate—and powerful—results. And what’s more, because of their simplicity, they are a great way for analytics neophytes to get their feet wet.

Specifically, we’re talking about four of the features in Analytics: clustering, keyword expansion, similar document detection, and concept searching.

 

Clustering

Clustering may be the best known of these features, but we often hear confusion about what do with clusters once they’ve been created. The most common use, of course, is to simply use them for preliminary exploration.

However, if you dig a bit further into the advanced settings, you have the opportunity to change your clusters from their default, nested setting, to a depth of 1, which creates flat clusters that look like this:

Easy Analytics Wins - Clustering 1

Flat clusters are particularly helpful for batching and reporting. For example, you can run a pivot to find out which search terms appear the most often in each of your clusters.

Keyword expansion

Keyword expansion is particularly well-suited for acquainting new users with conceptual analytics, largely because it has a proverbial foot in both the keyword and analytics universes. It’s a way to use conceptual analytics to help find better keywords. Some of the new keywords you find may be unlikely ones, which wouldn’t naturally come to mind when compiling a terms list. Keyword expansion is also a useful tool for QCing your terms list in order to find any straggling terms which you might have omitted accidentally.

Similar documents

The similar documents feature is the easiest way to find “more like these” documents. It’s used when you have already located documents and want to see what else is out there. Using this feature is as simple as right-clicking on a document in the viewer and selecting the ‘Find Similar Documents’ option. To maximize your results, it’s a good idea to modify the Search Results Pane view to give you more filtering, sorting, and tagging options:

Easy Analytics Wins - Similar Docs View

Concept Search

While the similar documents feature is used to “find more of these,” the concept search is used to find documents you’re not sure even exist. You can concept search off a portion of a document, off an external reference source, even off a fabricated document. This last feature can be a lot of fun for attorneys to experiment with. Simply write the best (or worst) document you can think of, and paste it into the search box. You don’t need to worry about exact word choice or order, because the system will search for documents containing the same concepts—a far more forgiving workflow than the typical constraints imposed by a standard keyword search.

 

All four of the above features are easy to use and provide immediate results. They are simple for attorneys and non-technical people to get a handle on, and help find items that might otherwise be overlooked if one was to rely solely on older methods.

For more advanced users, each of these features can even be used in tandem with your Assisted Review workflows to accelerate your projects and QC you work even more effectively.

For more information on text analytics and how you can make it work for your projects, please don’t hesitate to contact us. We’re happy to help.

 

Comments

Post a Comment

Required Field