Translating Big Data into Smart Data for Veterinary Epidemiology

VanderWaal, Kimberly; Morrison, Robert B.; Neuhauser, Claudia; Vilalta, Carles; Perez, Andres M.

In the coming decades, the greatest challenge in big data epidemiology will be to move toward creating value. Putting big data to work requires expanding our definition of the V's of big data to include three A's: accuracy, accessibility, and automation. While some may purport that the sheer quantity of big data means that inaccuracies in the data are washed out, this may not be true if issues with confounding, measurement error, and selection bias scale with sample size (5, 47, 62) . For example, data mining based on Google searches or Twitter may misrepresent the population at risk given biases in internet use. In addition, the value of extracting data from clinical records or diagnostic laboratory databases is dependent on the quality of record keeping and data entry. Thus, error checking and quality control should be incorporated into big data processing to ensure reliability (47) . Accessibility it also a critical challenge, encapsulating fundamental concerns related to data confidentiality and ownership (8, 14, 18, 47) , data engineering issues revolving around data structure Translating Big Data into Smart Data Frontiers in Veterinary Science | www.frontiersin.org