Visual analytics has arisen as a potentially game-changing advance in addressing big data challenges. The overwhelming complexity of big data has significantly challenged analysts as they attempt to envision the environment they seek to analyze, seeking to draw intelligent conclusions from observable data. Visual analytics offers a viable method of compensating for our cognitive limitations in processing vast amounts of data and our natural tendency for bias that ultimately cause us to ignore data, search in the wrong place, and ask the wrong questions. It empowers humans by offering automations that are capable of augmenting, and not interfering, with the human visual and reasoning processes, thereby taking full advantage of our innate skills in pattern recognition.
Visual Analytics in Cyber Security
While visual analytics is already extensively used in some areas, those are largely limited to research and academic environments. Therefore, major gaps exist between state-of-the-art visual analytics research and its application in operational environments. This lack of experience in the operational space has prevented the widespread adoption of tools and techniques whose effectiveness has already been experimentally proven. Without productive visualization techniques in operational environments, analysts—using outdated tools and techniques—are left ill equipped to glean information from today’s massive amounts of data. The deficiency is especially striking in fields such as cyber security where it cannot be known a priori which data are relevant. Cyber analysts require exploratory visualizations that empower them to drill down and investigate unknown threats without sacrificing their situational awareness of the environment at large. While current visualizations offer intuitive representations, they do not take advantage of humans’ innate exploratory capabilities. Improving visual analytics in an operational environment thus rests on first achieving an understanding of the human cognitive processes that are in play when visual representations are encountered. Accurately aligning the strengths of human and machine involves reaching across multiple social and scientific disciplines to achieve a better understanding of how concepts are stored and processed by the human brain. As a general rule, designers should improve visualization techniques, such as clutter reduction and highlighting, by drawing on current research in human perception.
Evaluating Visual Analytics Tools
Comprehending the analyst’s ability to perceive and understand data is critical to preventing the tools used from disrupting the analyst’s train of thought as he or she drills down and explores events. But an important question must be addressed: How do we accurately measure the effectiveness of a particular visualization technique or combination of techniques? Organizations are flooded with options for tools that claim to optimize interaction by capitalizing on human perception. But are those techniques effective for a particular organization’s specific data and work environment? Rather than blindly selecting one visualization method over another, organizations require a methodology to help them determine which visualization tools and techniques best align with their core functions and with the needs of their analysts. At Innovative Analytics & Training LLC (IAT) we understand the barriers to adopting new visual analytics tools. Visualizations are based on general patterns of cognition and perception, not designed specifically for operational end users. We help our clients explore the tools and options that are most suitable for their current analytic practices, workflow, and operational environment.
As researchers keep on expanding the functionality of visual analytic tools and techniques to explore, the demand for analysts capable of using them continues to increase. Training an analyst to use a visualization tool can be simple; training an analyst to think creatively poses a much greater challenge. Exploratory visualizations require a complete transformation in how analysts are trained, so that the process appropriately emphasizes creativity and prepares each analyst to fulfill his or her role as explorer. Rather than relying heavily on machines, training must focus on inspiring analysts to creatively explore those “what if” scenarios that lead to spontaneous insights. Often the most powerful visual analytics tools are very complex and demand advanced training. While each group generally possesses at least one expert user, his or her acquired knowledge of how best to leverage a tool is rarely passed down and institutionalized. Once a champion user moves to a new position, that knowledge is lost. At IAT we understand the vicious circle that can bedevil training and adoption: Analysts will take the time to learn a new capability only if they believe it adds an immediate and lasting value to their products; however, no value will be added if the analyst isn’t properly trained in using the tool. We work with clients to identify the visual analytics tools and methods that work best for each environment as well as provide the consistency and training required for long-term adoption.