Hypergraph: A Next-Generation Analytics Solution For Visualizing Data Insights

Tuesday, August 20, 2013 - 14:38

The Editor interviews Jerome Levadoux, Senior Vice President, Product Management & Marketing, Recommind.

Editor:  What technology trends are having an impact in the legal market today?

Levadoux: I see three interrelated technology trends. The first is Big Data, which is having a huge impact in every profession in which large amounts of data are handled daily. In the legal sector in particular, we are seeing the need to collect and process larger and larger amounts of information in real time so that professionals can gain insights and make decisions about that information with sufficient speed. In recent years, we’ve seen an absolute explosion in the volume of information that must be processed, and it’s overwhelmed anyone who is trying to understand or analyze this data manually or with traditional technologies.

The second fundamental shift is the advent of the cloud, which has had two major effects on business. For one, today’s corporations want to avoid the costly and time-consuming implementation and maintenance of software. They want to use software that is easily deployed from the cloud and doesn’t need to be maintained by an army of people.

In addition, use of cloud storage and services has caused a broad fragmentation of information – information that legal professionals will ultimately need to process. In years past, a company might host a few file shares, a customer portal and perhaps an email server, and almost all of the company’s communication and information would be handled by these. Today, people are storing and exchanging their information through the utilization of countless cloud depositories, using cloud storage and sharing tools, and communicating through external mail services. Consequently, for legal professionals trying to analyze information, to enforce the company’s governance or to collect information in the context of a lawsuit, their work has obviously become exponentially more complicated.

The third major trend we are witnessing is the consumerization of technology. Today’s technology has grown extremely user-friendly, and legal professionals don’t want to have to deal with very complex software. They have become accustomed to applications that are intuitive to use and visually appealing, and as a result they want software that will speak to them and immediately help them accomplish their tasks with minimal training.

Editor: What is Hypergraph?

Levadoux: An entirely new way of analyzing and visualizing large amounts of data, Hypergraph allows users to detect trends and patterns very quickly in digital information flows. Imagine someone who spends a lot of time analyzing communication data or information flows – i.e., email, social media, or any information source that is relevant in an information governance, compliance, auditing or investigation context. Hypergraph is a rich graphical environment through which a user can see relationships between pieces of information within large, complex data sets. Unlike other technologies with scalability limitations, Hypergraph can handle very large amounts of data, and can do so rapidly.

Previously, when a user had to analyze data for e-discovery, he would be forced to manually search through thousands of emails and other sources of information one at a time. But a lot of very relevant information cannot be encapsulated in one single document or one isolated email. Hypergraph enables the user to analyze a large array of documents at once and identify relationships between those documents, for instance, who is speaking with whom, the subject of their conversations, and when the communication took place. With Hypergraph, you can produce a histogram (or timeline) for viewing date-based information in order to follow emails over time and see the volume of communication between individuals.

By looking at information in the aggregate all at once, you may begin to detect patterns that you would never have been able to discover had you looked at each document separately. By stepping back from the “noise” of large data sets, you can detect trends, relationships and patterns in the information that may tell a particular story.

Editor: How does Hypergraph improve upon similar technologies in the marketplace? Does it address the challenges implicit in the legal technology trends you spoke of earlier?

Levadoux: I believe that Hypergraph offers several critical advantages over more traditional technologies. The first is scalability: Hypergraph can process very large amounts of information in a very short period of time. I mentioned earlier that the ever-growing volumes of information in today’s enterprises can easily contain billions of pieces of information. Most other solutions cannot scale at an acceptable rate or have a slow response time that inhibits productivity. The ability to identify and analyze billions of items in real time is a significant technological leap that we are very proud to offer to the market.

The second advantage is integration. Our solution is built on our CORE® platform, and this data management analytics platform is the underpinning of everything we do at Recommind. The same platform also serves as the basis for our search, email, e-discovery and  information governance products. Hypergraph is integrated into the CORE platform, resulting in a completely unified, out-of-the-box solution for data extraction, data analytics and visualization. In order to accomplish the same task without our technologies, you would have to hire external system integrators to stitch together a host of different technologies to achieve the same or similar results, which would cost you significantly more time and money.

Third, Hypergraph enables users to go beyond the pure analysis of data and to take action. Recommind wanted the analytics phase to be directly integrated into the workflow such that while looking for information, a user can almost instantly detect patterns and trends – and immediately act on them. This is especially relevant in the context of e-discovery. For example, while conducting a search, a legal professional finds a group of relevant documents among a corpus of data. With a single click in Hypergraph, the user is able to tag all of the documents that have been isolated. This type of capability used to be impossible! Hypergraph automates that entire process, so once relevant information is found during analysis, a user can click and be done.

Editor: Are different visualization schemes available in Hypergraph?

Levadoux: There are several deductible visualizations available. First, a dashboard of controls enables the user to refine the navigation and interact with data. A large array of filters allows the user to select information based on specific people, locations, dates and times, or domain names, etc. You can also select a particular path of communication, for example, whether it was emails sent inside or outside the company firewall.

The Hypergraph search box allows a user to conduct a simple keyword search or a very complex Boolean search, or to search for specific concepts in the corpus of data that’s being analyzed. Meanwhile, other control options will narrow and visualize the scope of information for a user.

In addition, there are three key kinds of visualization. One of them is the histogram, which provides a view of how a particular set of data has evolved over time. The second is a graphical representation of entities and, very importantly, the interrelationships between them. An entity might be a person, a topic or one of any number of objects or events. Third are the smart charts. Those charts can be pie charts or bar charts that show you information in its purely quantitative aspect.

Below please find an example of Hypergraph’s visualization capability (click to enlarge):

Hypergraph provides a new way of analyzing and visualizing data to discover patterns and relationships in digital information flows.  By combining elements of machine learning, graph analysis, and a rich visual environment, Hypergraph allows people to quickly see the big picture, gain insights and take action.

Editor: How have legal professionals managed piecing together information flows up to this point?

Levadoux: With a lot of manual work! Imagine analyzing a thousand emails and finding relationships between those emails, let alone trying to discern a pattern of who has been talking to whom, who has been the most conversant, or whether some people are hidden from the conversation. This process is highly tedious and painstakingly long.

Editor: Who within a company should be using Hypergraph?

Levadoux: There are several use cases. As a very horizontal analytics offering, Hypergraph can be used for a wide range of use cases. Not surprisingly, some of the immediate use cases are around corporate e-discovery. As people analyze information related to a specific case, they can use Hypergraph to cull digital communication from the search, to find relevant documents and to identify people involved in a specific conversation. Hypergraph may also help an analyst discover patterns of unique conversation threads between individuals, and to conduct a more robust custodian analysis. It can also shine a light on anomalies, such as missing or deleted communications.

Based on the feedback we’ve received, we’ve learned that Hypergraph will also be very useful for corporations conducting internal investigations. An HR department may use it to assist in the investigation of a sexual harassment case, or a compliance officer may use it to aid in an anticorruption case.

The third use case is the most widely applicable to corporate regulatory response. Organizations in regulated industries must be able to respond to an agency’s information requests in a timely manner; they want to quickly identify all related content they have to provide to the agency in response to a specific inquiry.

Editor: How does Hypergraph compare to other visualization technologies on the market now?

Levadoux: When we looked at the competition, we found a set of pure visualization vendors. As I was saying earlier, these are not integrated with a data management platform, so one would have to cobble together a few technologies to make it work. Most companies don't want to take on that costly, time-consuming task.

Hypergraph is extremely fast, compelling and easy to use. Anyone using Hypergraph can be productive in minutes. It allows you to ask questions as naturally as you would in Google or Excel and is extremely user-friendly.

Editor: So, does the Hypergraph offering come separately or does it reside within your newest solutions? How may a customer acquire it?

Levadoux: Hypergraph is a separate product that can be licensed to any customer, whether it’s an existing or a new customer at Recommind. As we introduce this product, one of our target areas is our existing Axcelerate On-Demand customers – our on-demand e-discovery customers – because we believe that Hypergraph is a very simple add-on that we can turn on very quickly and provide immediate value, in addition to the e-discovery or the review capabilities that they are already using. But it’s also licensable as a stand-alone product.

Editor: Are there any concerns about keeping sensitive information in the cloud? What are the security risks?

Levadoux: Every company in every industry sector, and I would say in every country, has slightly different perceptions and approaches to storing information in the cloud. We have a very large number of customers, and we see among them a clear trend to conduct e-discovery in the cloud. That’s by far the fastest-growing segment in our business at Recommind.

We already have some of the largest and most prominent global corporations entrusting us with their information – much of it confidential and sensitive, mission-critical information – in our e-discovery cloud.

So obviously these companies have already made that mental step to store information in our cloud, and Hypergraph simply gives them another way to view and analyze it. Beyond that, every corporation is unique. Of course, different kinds of organizations – be they public sector, banks or manufacturing companies – have different sensitivities to cloud storage and cloud-based solutions. Many companies in Europe are more nervous about moving their information to the cloud than companies in the U.S. because of privacy and security concerns. Overall, we are seeing that more and more companies are becoming comfortable with utilizing the cloud.

Editor: Was there anything else you’d like to add?

Levadoux: I suppose I would like to say that Hypergraph, with its rich graphics environment fully integrated into the Recommind CORE platform, represents the next generation of analytics. Recommind created Hypergraph in response to people’s need to detect patterns and to gain new insights from large amounts of communication information flows. I believe we have gone even further by building a solution that allows a user to take immediate action during workflow, greatly increasing his or her productivity.

Editor: I know our readers would like to find out where to find more information.

For more information on Recommind’s latest product announcement, visit www.recommind.com/products/hypergraph.

Please email the interviewee at jerome.levadoux@recommind.com with questions about this interview.