Unit 2 · Data

Bit ● (core concept) — Shorthand for binary digit: a 0 or 1. Bits are the lowest-level components of any digitally represented value.

Byte ● (core concept) — A group of 8 bits, the basic building block used to assemble larger data representations.

Abstraction ● (core concept) — The process of reducing complexity by focusing on the main idea: hiding details that are irrelevant to the question at hand and bringing together related, useful details so one can focus on the idea.

Analog data ● (core concept) — Data whose values change smoothly over time rather than in discrete intervals, such as the pitch or volume of music, the colors of a painting, or the position of a sprinter during a race.

Digital data ● (core concept) — Data stored as discrete values built from bits. Using digital data to approximate real-world analog data is itself an example of abstraction.

Sampling ● (core concept) — A technique for approximating analog data digitally: measuring the values of the analog signal at regular intervals, called samples, and determining the exact bits required to store each sample.

Binary number system ● (core concept) — The base-2 system, using only combinations of the digits 0 and 1. A digit's numeric value equals its bit value (0 or 1) times its place value, where each position's place value is the base raised to the position's power, numbered from 0 at the rightmost position.

Overflow (fixed-bit integers) ● (core concept) — In many programming languages, integers are represented with a fixed number of bits, which limits the range of values and the operations possible on them; exceeding that range can produce overflow or other errors. The same sequence of bits can represent different data types in different contexts.

Round-off error ● (core concept) — Error arising because real numbers are stored with a fixed number of bits, so their range and precision are limited; some real numbers are stored only as approximations.

Data compression ● (core concept) — Reducing the number of bits used to store or transmit data. The amount of size reduction depends on how much redundancy the original data contains and on the compression algorithm applied.

Lossless compression ● (core concept) — A compression approach that usually reduces the number of bits stored or transmitted while guaranteeing complete reconstruction of the original data. Fewer bits does not necessarily mean less information.

Lossy compression ● (core concept) — A compression approach that can significantly reduce the number of bits stored or transmitted, but only allows reconstruction of an approximation of the original data.

Choosing a compression algorithm ● (core concept) — Lossless algorithms usually shrink data less than lossy ones. Choose lossless when quality or the ability to reconstruct the original is maximally important; choose lossy when minimizing data size or transmission time is maximally important.

Information ● (core concept) — The collection of facts and patterns extracted from data. Data provide opportunities to identify trends, make connections, and address problems.

Correlation and causation ● (core concept) — Digitally processed data may show a correlation between variables, meaning they move together. A correlation does not necessarily indicate a causal relationship; additional research is needed to understand the exact nature of the relationship.

Metadata ● (core concept) — Data about data, such as an image's creation date or file size. Metadata are used for finding, organizing, and managing information and can make data sets more effectively usable; changes or deletions to metadata do not change the primary data.

Data cleaning ● (core concept) — A process that makes collected data uniform without changing their meaning, such as replacing all equivalent abbreviations, spellings, and capitalizations with the same word. Nonuniform collection, for example free-text entry by different users, makes cleaning necessary.

Data bias ● (core concept) — Skew in conclusions created by the type or source of the data being collected. Bias is not eliminated by simply collecting more data.

Large data sets ● (core concept) — Data sets so large they are difficult to process on a single computer and may require parallel systems. A system's scalability, its capacity to grow in size and scale to meet new demands, determines how such data can be stored and processed.

Data processing with programs ● (core concept) — Using programs to process data and acquire information. Programs are applied in an iterative and interactive way so users can gain insight and knowledge, and programmers can filter and clean digital data with programs.

Data transformation operations ● (core concept) — Standard processes for extracting or modifying information from data: transforming every element of a data set, filtering a data set, combining or comparing data, and visualizing a data set through a chart, graph, or other visual representation.

Search tools and data filtering ● (core concept) — Tools for efficiently finding information and recognizing patterns in data. Search tools speed up finding information; filtering systems help surface patterns; programs such as spreadsheets help organize data and find trends.

Combining data sources ● (core concept) — Merging, clustering, or classifying data from multiple sources to gain insight no single source could support. A single source often lacks the data needed to draw a conclusion, and patterns can emerge once data are transformed by programs.