Data Quality Dimensions
You'll learn the six fundamental dimensions used to measure and assess the quality of data in production pipelines.
The Core Idea
Data quality isn't a simple yes-or-no judgment—it's multidimensional. When monitoring pipelines (as we've been doing), we evaluate data against six key dimensions: accuracy, completeness, consistency, timeliness, validity, and uniqueness. Each dimension addresses a different way data can degrade or fail to meet requirements.
Think of these dimensions like inspecting a bridge: you check structural integrity (accuracy), whether all parts are present (completeness), that design specs match across sections (consistency), that it's maintained on schedule (timeliness), that materials meet codes (validity), and that no duplicate bolts were accidentally installed (uniqueness).
The Six Dimensions
Accuracy: Does the data reflect reality? Are measurements correct? Example: Customer ages matching birth dates.
Completeness: Are all required fields populated? Example: No missing email addresses when emails are mandatory.
Consistency: Does the same data match across systems or time? Example: A customer's name spelled identically in billing and shipping tables.
Timeliness: Is data available when needed? Is it fresh enough? Example: Real-time stock prices updated within seconds, not hours.
Validity: Does data conform to defined formats, ranges, or business rules? Example: Dates formatted as YYYY-MM-DD, not as text strings.
Uniqueness: Are there no unintended duplicates? Example: Each customer appears only once in the master customer table.
Applying Dimensions in Practice
When building data quality checks (lesson 1859), you're actually testing these dimensions. Your pipeline monitors measure violations across these categories, alerting when thresholds are breached.
Key Takeaway: Data quality is measured across six dimensions—accuracy, completeness, consistency, timeliness, validity, and uniqueness—each capturing a distinct way data can fail to meet requirements.