Big Data

From Data Science
Jump to navigation Jump to search

Many different sources have tried to define what they mean by "Big Data", so rather than providing a single definition here, we summarize a few of them.


Oracle defines big data as data which shares the following four features:

  • Volume: machines are capable of generating easily petabytes of data in a day, so the core feature of big data is that the data sets are... big.
  • Veleocity: not only are big data datasets large, they grow quickly. Consider a social network or set of physical sensors - there is not just an existing dataset to analyze, but a constant, fast stream of new data constantly being added.
  • Variety: big data often includes efforts to combine disparate data sets from various sources, and draw connections between them. For example, a sensor network might have an entirely new class of sensors added, to measure a different parameter, and this data can be utilized side by side with existing data.
  • Value: the point of big data isn't just to accrue massive amounts of information - it is to extract useful, valuable insights from that data. The data itself holds an intrinsic value which can be used, leveraged, and put to work. The challenge is then extracting the valuable information from within a much larger data set.