CompTIA Data+ key terms

Data organized into a fixed, predefined format such as rows and columns in a table, where each field holds a known type of value. It is easy to search, sort, and query with standard tools.

Data that has no predefined model or fixed format, such as free text, images, audio, and video. It usually requires specialized tools or processing before it can be analyzed at scale.

Data that carries some organizing markers, like tags or key-value pairs, but does not fit neatly into rows and columns. JSON and XML documents are common examples drawn from web and log sources.

A column or set of columns in a table whose values uniquely identify each row. No two rows may share the same key value, and the key cannot be left empty.

A column in one table that references the primary key of another table, linking related records together. It enforces referential integrity so a child row cannot point to a parent that does not exist.

The process of organizing tables to reduce duplicate data and avoid update anomalies, typically by splitting information into related tables. It usually moves a database toward third normal form.

A central repository that stores cleaned, integrated data copied from many source systems, structured for reporting and analysis. It typically holds historical data and supports complex queries rather than day-to-day transactions.

A storage repository that holds large volumes of raw data in its native format, including files, logs, and streams, until it is needed. Structure is applied at read time rather than load time.

Extract, transform, load: a pipeline pattern that pulls data from source systems, cleans and reshapes it, then writes it into a target such as a warehouse. Transformation happens before the data is loaded.

The practice of exploring large datasets with statistical and machine learning techniques to uncover patterns, relationships, and trends that are not obvious on the surface.

A machine learning approach that trains a model on labeled examples, where the correct answer is already known, so it can predict labels or values for new, unseen records.

A machine learning approach that works with unlabeled data, looking for natural groupings, structures, or relationships without any target answer to guide it.

A supervised technique that assigns records to one of a set of predefined categories, such as flagging a transaction as fraudulent or legitimate, based on patterns learned from labeled examples.

An unsupervised technique that groups records so that items inside a group are more similar to each other than to items in other groups. No predefined labels are used to form the groups.

A technique that finds items or events frequently occurring together in transaction data, producing statements like customers who buy one product often also buy another. It underpins market basket analysis.

A modeling problem where a model learns the training data too closely, including its noise, so it performs well on data it has seen but poorly on new data.

Summary measures that describe the main features of a dataset, such as counts, averages, ranges, and spread, without drawing conclusions beyond the data itself.

A measure of how spread out values are around the mean. A small value means most points sit close to the average; a large value means they are widely scattered.

A statistical measure of how strongly two variables move together, ranging from minus one to plus one. It shows association but does not by itself prove that one variable causes the other.

A technique for modeling the relationship between one or more input variables and a numeric outcome, often used to predict values or estimate the size of an effect.

A formal procedure for deciding whether observed results are likely due to chance. It compares sample data against a null hypothesis and yields a probability used to accept or reject that hypothesis.

The probability of seeing a result at least as extreme as the one observed if the null hypothesis were true. Small values suggest the observed effect is unlikely to be random.

A range of values, calculated from a sample, that is likely to contain the true population value. It is usually reported alongside a confidence level such as 95 percent.

An experiment that splits users or subjects randomly into two or more groups, applies a different treatment to each, and compares outcomes to see which variation performs better.

The practice of presenting data as charts, graphs, or maps so that patterns, outliers, and comparisons can be understood quickly by an audience.

A chart that groups numeric data into bins and shows how many values fall in each bin. It reveals the shape, center, and spread of a distribution.

A chart that places one variable on each axis and draws a point for every record, making it easy to spot relationships, clusters, and outliers between two numeric measures.

A chart that shows a distribution through its median, quartiles, and whiskers, with points beyond the whiskers flagged as potential outliers. It is useful for comparing groups side by side.

A single screen that combines several charts and summary figures, usually refreshed on a schedule, so viewers can monitor performance at a glance.

Key performance indicator: a measure chosen because it directly reflects progress toward an important business goal, and is tracked consistently over time against a target.

The craft of pairing a visual with a clear narrative so an audience understands what the data shows and what action is being recommended, rather than just seeing numbers.

The set of policies, roles, and processes that define who is allowed to use which data, for what purpose, and how quality is maintained across an organization.

The degree to which data is accurate, complete, consistent, timely, and fit for its intended use. Poor quality shows up as errors, duplicates, and missing values.

Data about data: information such as field names, data types, owners, source systems, and refresh dates that describes a dataset and helps people find and trust it.

A record of where data came from, how it moved and changed through each system or transformation, and where it ended up. It supports troubleshooting, auditing, and impact analysis.

An organized inventory of an organization's datasets, searchable by name, description, and metadata, so analysts can discover what data exists and who owns it.

Personally identifiable information: data that can be used alone or combined with other details to identify a specific individual, such as names, addresses, or government ID numbers.

A technique that replaces sensitive values with realistic but altered substitutes, so testers and analysts can work with the data without seeing the real personal details.

The discipline of creating and maintaining one trusted, consistent version of core reference data, such as customers, products, or suppliers, shared across all systems.

A person responsible for the day-to-day care of a data domain, including defining business terms, resolving quality issues, and approving changes to how the data is used.