Data quality
Data quality is the degree to which data are accurate, complete, consistent, timely, valid, and suitable for their intended uses.
Data accuracy: The degree to which data values correctly represent the entities or events they describe. Accuracy is one dimension of data quality, but accurate values can still be incomplete or mistimed.
Data profiling: The examination of data to summarize structure, content, patterns, and anomalies. Profiling exposes missing values, unexpected distributions, and rule violations that signal quality issues.
Data governance: The framework of decision rights, responsibilities, and policies for managing data. Governance assigns ownership for quality standards, stewardship, and escalation of defects.
Garbage in, garbage out: The principle that flawed inputs can produce flawed outputs in computation and analysis. It captures how downstream processing cannot reliably repair defects inherited from source data.
Data completeness: The degree to which required records, fields, or values are present in a dataset. Missing records or fields lower quality even when every recorded value is accurate.
Data validation: The checking of data against specified rules, formats, or constraints. Validation detects violations of known requirements during entry, transfer, or processing.
Master data management: The discipline of creating and maintaining consistent core records shared across an organization. Reliable customer, product, and supplier records depend on resolving duplication and inconsistency.
Measurement error: The difference between an observed measurement and the quantity it is intended to measure. Systematic or random measurement errors directly reduce the accuracy of recorded data.
Data consistency: The degree to which data agree across fields, records, systems, or representations. Conflicting values across systems reveal a quality problem distinct from inaccuracy in any single record.
Data cleansing: The process of detecting and correcting or removing inaccurate, incomplete, or inconsistent data. Cleansing changes problematic records after quality defects have been identified.