Principles and Practice of Big Data, 2nd Edition

16

Data Reanalysis: Much More Important Than Analysis

Abstract

It is commonly assumed that studies conducted with large sets of data provide definitive answers; the idea being that if you have all of the available data at your disposal, you cannot miss coming to a correct conclusion. As it happens, large data sets seldom if ever yield straightforward solutions. There are just too many variables, too many ways of approaching problems, and too many ways of interpreting results. In this chapter, we will see that the first analyses of Big Data are nearly always wrong. Not to worry; if the data has been properly collected, organized, annotated, and archived, we can reanalyze the data until it begins to make sense to us. This chapter discusses the ...