Make way Palantir and IBM – Linkurious makes access to graph database visualization a snap. No developers or data scientists required.
More than a decade after the official “dawn” of Big Data analytics, there is general agreement that raw data findings alone are but the launchpad to the true end game of analytics: visualization that simplifies understanding of trends, coming events and perhaps most importantly – relationships.
In the ISS sphere, companies such as Palantir and IBM have made a mint providing visualization software that graphically depicts the tail end of Big Data to provide “actionable intelligence” to law enforcement and government agencies. Good as they are, both are dwarfed by the world’s most popular graph database solution Neo4j, developed by Neo Technology in 2007 and today used by leading enterprises including eBay, Walmart and Pitney Bowes to understand and meet customer needs. Neo4j, like Palantir and IBM i2, visualizes data faster, more accurately and far more economically than conventional database tools such as SQL and NoSQL.
At the recent ISS World Middle East event in Dubai, GraphAware demonstrated how the Neo4j database may be used to “untangle the criminal web using Neo4j.” GraphAware makes tools for software geeks that speed the process. What he left out: Neo4j is complex. Using it is a job not just for developers and certainly not for the average end user, but for data scientists. While large government agencies often do have large staffs of algorithm-mulling math wizards on board, it’s far less less likely to find such lofty types at the typical LEA.
For that reason, startups like Linkurious have emerged with tools specifically designed to make graphbase searching and queries simple to use. Here we’ll take an in-depth look at priceless value of Neo4j, and how Linkurious steps in to make the visualization easy.
Neo4j and GraphAware – Accelerating Contextual Intelligence
Understanding what makes Neo4j special requires wrapping one’s mind around a whole new approach to database management systems (DBMS) and relational database management systems (RDMBS). Conventional DBMS and RDMBS, and for that matter Big Data systems, hold data in language and numerics. The end result of queries and analytics: numbers and text, for example, on types and count of threats presented in columns and tables.
The graph database has a different and expanded mission: to show the relationships between data findings. To do so, the graph database uses “graph structures” – lines, edges and arcs – to represent and store data for semantic queries, i.e., queries that reveal context and associations. Each image represents specific items of data stored as a graph, not as text and not numerically, but as “nodes” (circles) connected by “edges” (lines).
The relationships in a graph database are inherently linked together – nodes to edges to other nodes, etc. As a result, semantic queries on a specific topic can instantly produce a hierarchical structure of graphs in a single image. Color-coded for different features, the graphs reveal not just bare data but the relationships between what might otherwise seem disparate findings.
The graph database is light years removed from the traditional relational database. In a SQL database, for example, the user must tap the “join” function, which then works through the process of combining columns and tables based on common values in each. It is a slow and cumbersome process. In contrast, the data in a graph database is already linked in a “nodes and edges” diagram, and is technically more streamlined.
In a commercial setting, a graph database might show a retail store how female shoppers, age 25 to 30, have given preferences for purchases of specific merchandise, and a predilection for complementary purchases on or around set days and times, and more likely to spend when the price is at or below a set level, or taking advantage of loyalty discounts. Neo4j might also reveal friends in common, skills in common, whether targets live close or work together, and the degree of personal closeness or separation.
In the ISS arena, the graph database might be used to show the relationships between criminal or terrorist targets of a certain age, education, livelihood, social media preference, communications with others, and similarity in profile to individuals not previously known to law enforcement, but definitely linked.
Raising the Bar on Data Value
The business case around Neo4j is “building sustainable competitive advantage” via the ability to fathom, leverage and profit from customer relationships. In other words, Neo4j boosts the value of data by pulling more insights from it than is possible with conventional RDBM systems than NoSQL. Because the graph database is faster and more economical, costs drop, as well. For all these reasons, Forrester Research predicts that by the year 2020 some 25 percent of enterprises will have migrated to graph databases.
The lesson is not lost on the ISS community, which has historically shown interest in the graph database approach of Palantir and IBM, and is now seeing the light in open source Neo4j.
Users can get a boost from niche products by GraphAware, which provides consultancy and six different software products that can help ease the way into Neo4j for parties interested via recommendation engines, time instants linked to events, Universally Unique Identifiers (UUIDs) that lock in results of newly created nodes in a graph, training in Cypher, Neo4j’s query language, or how to use Neo4j with PHP, the common scripting language for HTML.
Just one problem, as you may have already surmised from the preceding paragraph: Both Neo4j and GraphAware are not kid stuff – they’re expressly designed for developers with foreknowledge of Java. That description might not apply to every law officer or government intelligence officer eager to reap the benefits of graph database management for visualization.
Voila! Enter France’s Linkurious
As graph databases began to gain momentum, Paris-based Jean Villedieu saw an opportunity. While the market was flocking to Neo4j, many enthusiasts found they were hitting a wall. They needed a PhD in Data Science to enter the world of relationship visualization made possible by Neo Technologies and GraphAware. Villedieu has described the problem in vivid terms:
“The problem with graph databases is you don’t have tools that can make it easy for end users to extract information. In relational databases there are tools like Tableau and Qlikview, and all the business intelligence solutions that do a great job of making sense of relational data. But there’s nothing for graph data. You only have tools that are designed for scientists.”
With four partners, Villedieu in 2013 created a new venture to solve this problem – Linkurious. With the company’s eponymous software solution, users can access visualizations from a graph database like Neo4j simply by keyword search. Results are displayed in “nodes and edges,” the same as a developer or data scientist would see after applying their skills Neo4j directly or with help from GraphAware.
Design your own visualization tools just to access graph data in Neo4j? Forget it. Linkurious is “Everyman’s” entry point to seeing brilliant, insightful visuals that shed new light on investigations and intelligence activities. Linkurious works with popular graph databases including Titan, Datastax Enterprise Graph, Franz Allegrograph and Neo4J. No data scientists or developers required.
Insider Surveillance rating for Linkurious: 5 Stars.