Message from the Director
Big Data: Old Hat and Forte of the VACCINE Team Part 2
Visual analytics can help solve several of the problems that are not addressed by the proposed mainstream big data analytics techniques. As noted by David Brooks in his New York Times article on February 18, 2013, big data analytics can’t handle qualitative, fuzzy, and social data well (e.g., trust, preferences). Humans are much more adept at reasoning with such data. Similarly, incomplete context creates significant problems for automated analytical algorithms and for most complicated, real-world situations, the available data sources normally are not complete and don’t have all the information that goes into the final decision making as do the human decision makers. These examples are exactly where visual analytic approaches are much better solutions since visual analytics capitalizes on the best and complimentary abilities of both components of the human-computer decision-making process through iterative, interactive visual interfaces to leverage and supplement the powerful cognitive capabilities of the human user. Our visual analytic solutions provide relevant information for the decision maker to supplement and enhance their decision making ability, whether they are making operational allocation decisions for the U.S. Coast Guard or a law enforcement officer on the street.
Automated analytical algorithms employing statistical analysis with large data sets can find many statistically significant events, correlations, and factors, but without the appropriate context and deeper understanding of the problem/phenomena under study, the majority of these may be completely irrelevant to the decision maker or scientist. Therefore, this additional irrelevant information leads not only to more work by the decision maker to tease out useful information, but leads to less trust in the usefulness and results from the algorithms. By having the decision maker, officer, and scientist/engineer interactive drive and interrogate the analysis and presenting the information in a relevant context, visual analytic approaches overcome this issue.
As Brooks also notes, visual analytic approaches are also much better at tackling big, complex, multifaceted, multiparameter problems (e.g., grand challenges) because they help organize information, and knowledge used in the solution process instead of try to solve the problem with automated algorithms alone when there is incomplete information involved (e.g., incomplete context, additional factors, historical information, unquantified dependencies, hard to predict public reactions).
As you can see, our visual analytic approach has many advantages since the goal is to empower the scientist, decision maker, engineer to be more effective in their task by providing additional, timely, guided, relevant knowledge for their task. These solutions do require close interaction with our partners to ensure we are providing relevant, useful knowledge and we have many years of experience in forming these partnerships to help solve difficult problems. We are always looking for new partners and new challenges, so please contact us if you think we can be of assistance.