Data literally means things given, and it can outlast its disks
The Latin datum meant something given, and English borrowed the plural in 1645. Three centuries later, in 1946, it came to mean computer information. Now science produces so much of it that archivists face an awkward truth: a paper book can outlive the hard drive or optical disc holding the same results.
At its simplest, data is a collection of values, discrete or continuous, that carries information; a single value is a datum or data point. It is gathered by measuring, observing or querying, then usually cleaned before analysis by dropping outliers and fixing obvious entry mistakes. Field data comes from uncontrolled real-world settings, while experimental data is produced under controlled conditions. Price indices, literacy rates and census counts are everyday examples.
A common hierarchy places data at the bottom, information above it and knowledge at the top. The measured height of Mount Everest in a database is a datum; a book on the mountain's geology is information; a climber's guide to reaching the summit is knowledge. Some thinkers argue this ladder runs the wrong way. Beynon-Davies separates the terms using signs: data is a string of symbols, and information appears once those symbols point at something. How informative data is depends on how unexpected it is, a quantity measured by Shannon entropy.
Computers store data in two broad styles. Analog machines use a physical quantity such as a voltage or position, while digital ones use symbols from a fixed alphabet, most often binary zeros and ones. A program is itself data that can be read as instructions, and in Lisp code and data are nearly indistinguishable. Metadata describes other data; the library catalogue is the classic case. Marketers add their own taxonomy, from zero-party data that customers volunteer to third-party data aggregated by outside firms.
Grammar is unsettled. Used as a general synonym for information, as in big data, the word behaves like a singular mass noun; in the sciences, referring to sets of measurements, it often stays plural, and APA style in its 7th edition insists on the plural. Meanwhile big data at the petabyte scale has pushed the growth of data science and machine learning, and fields like genomics and astronomy still lack a satisfactory way to preserve their output across centuries.
Source: Data
