This blog post is an excerpt from GovLoop’s recent guide, “Putting Data Analytics at the Forefront of Your Agency.”
While many agencies are beginning to use analytics to understand events impacting their organization, those system-centric approaches often miss relevant connections between other objects, people, places or events. Without that context, agencies are forced to work with an incomplete view of their organization to inform decisions.
To gain a more holistic picture of what’s happening at your agency, Roman Chanclor, Solution Sales Specialist, and Clint Green, Director of Advanced Analytic Strategies and Development at ViON, suggested taking an entity-based approach to data analytics. ViON is a leading provider of solutions to help organizations make the most of their data.
Entity analytics is an incremental context accumulator for detecting like and related entities across large, sparse, and disparate collections of data, including both new and old data, small and big data environments, to perform analytics on events, people, things, transactions, and relationships. In other words, entity analytics uses data from multiple sources and assembles a single story by making connections between those disparate points.
This approach is critical to many agency operations. Consider background investigations, for instance, which most agencies have to perform to hire employees or grant visas. Entity analytics can leverage big data analytics to examine an individual’s background from all perspectives. Social media profiles, aliases, public associations, and other personal traits can be correlated with demographic and government information like driver’s license and passport details. Together, those details can create a more complete picture to aid in the investigation.
“When you start to connect individual events to people, to locations, and to other entities, you start to understand them in a much broader concept,” said Green.
But for organizations without advanced analytics tools, achieving this broader view can be a challenge. Determining correlation and causality between entities requires an in-depth understanding of how those objects and events connect on operational and technical levels across the organization. However, organizational and technical siloes prevent IT personnel from easily combining and analyzing data from across the enterprise.
That data is also often a combination of structured and unstructured information, making it even more difficult to consolidate and compare. For instance, one person may be defined by a number of different identifiers, such as a social security number, a social media handle, a location, a name and even a nickname. To track that person across an enterprise, those disparate data points have to be reconciled and then analyzed. Then they have to be related to other entities, which might also have numerous different identifiers.
Entity analytics is a complex process. IT staff commonly attempt to overcome these challenges through manual methods. However, that approach can’t match the explosive rate of data growth today.
Nevertheless, agencies shouldn’t be dissuaded from pursuing this approach to analytics. “It’s important to decompose all the effort that goes into entity analytics so people can understand that it’s approachable. There isn’t a huge barrier to entry if you use the technologies available,” Green said.
“There’s no reason to leave that context on the table, for fear that it might be too complicated a task,” agreed Chanclor.
The right tools can vastly simplify and empower an agency approach to entity analytics. Automated solutions can quickly sort and analyze data ¬– even unstructured and disparate information ¬– without taxing IT personnel. For instance, entity analytics can be applied to cybersecurity and fraud prevention operations with ViON’s DataAdapt Threat Detect. That solution ingests data from available sources and performs analytics to detect intersections in the data for cyberthreat and fraud detection, continuously learning with each new set of data.
The platform also accomplishes entity resolution. “Our approach uses a probabilistic engine that applies a statistical understanding of entities,” said Green. “Given what is known about an entity already, data can be disambiguated, de-duplicated, or in other ways refined to provide a more accurate picture of the real world situation.”
Once IT personnel know who and what is happening on their network, they can begin to understand how different entities interact to form a more complete view of their organization. They can then use that understanding to inform proactive cybersecurity, fraud detection, or other mission strategies.
In the face of constantly evolving technology, it’s imperative for government to move beyond manual, reactive processes of data identification. Entity analytics can help IT personnel move their agencies forward with automated tools that consolidate and make the most of disparate government data.