BCA Data Science Big Data Characteristics — practice questions

1. The framework's definition is as short as it gets: "Velocity refers to the rate of data flow." 2. The body section adds what follows from it: "Velocity is a measure of the rate of data flow", and high-velocity systems "have been described as streaming data". 3. The architectural consequence is that data in motion "must be handled in a very different way than data at rest", tending to resemble event-processing architectures and often being processed in memory without being stored. 4. The size of the dataset is volume, data from multiple repositories, domains or types is variety, and changes within a dataset is variability. All four are defined within a few lines of each other, so the exact wording is the only reliable discriminator. _Source: NIST SP 1500-1r2 "NIST Big Data Interoperability Framework: Volume 1, Definitions" (Oct 2019), p.8 section 2 "Terms and Definitions"_

1. The definition reads: "Variety refers to data from multiple repositories, domains, or types." 2. The body section restates it as "the need to analyze data from multiple repositories, domains, or types", and explains how this used to be handled: by finding features that allowed datasets to be aligned and fused into a data warehouse. 3. The framework notes that automated fusion now "relies on semantic metadata, where the understanding of the data through the metadata allows it to be integrated". 4. Data arriving faster than it can be stored is velocity. Individual values changing is closer to what the framework calls volatility. Size beyond one machine is volume. 5. Variety is singled out for a reason worth remembering: the framework says volume and velocity make analytics faster and cheaper, but "it is the variety of data that allows analytic results that were never possible before". _Source: NIST SP 1500-1r2 "NIST Big Data Interoperability Framework: Volume 1, Definitions" (Oct 2019), p.8 section 2 "Terms and Definitions"_

1. The framework introduces a third category between the two familiar ones. 2. It writes: "Semi-structured data refers to datasets in formats such as XML (eXtensible Markup Language) or JavaScript Object Notation (JSON) that has an overarching structure, but with individual elements that are unstructured." 3. So the defining feature is the split: structure at the level of the document, no structure inside the individual elements. 4. That is why neither of the two simple labels fits. Structured data is defined as having "a predefined data model or is organized in a predefined way", and unstructured data as having neither, with text, image and video as the examples. 5. The framework's broader point is that the Big Data paradigm shift "has increased the emphasis on extracting the value from unstructured or relationship data, and on different engineering methods that can handle these types of datasets more efficiently". _Source: NIST SP 1500-1r2 "NIST Big Data Interoperability Framework: Volume 1, Definitions" (Oct 2019), p.12 section 3.2.5 "Structured and Unstructured Data Types"_