Data Quality Dimensions
What you'll learn: The five critical ways to measure whether your data is trustworthy and usable.
Why Data Quality Matters
Before you can analyze data or build models, you need to know if your data is actually good. Just like food has expiration dates and quality grades, data has measurable quality dimensions that tell you whether it's reliable enough to use.
The Five Key Dimensions
Accuracy
Does the data reflect reality? If a database says someone is 200 years old, that's probably inaccurate. Think of accuracy like a target: how close are your data values to the true bullseye?
Completeness
Are there missing pieces? If customer records lack email addresses or purchase histories have blank fields, your data is incomplete. Imagine a puzzle with missing pieces—you can't see the full picture.
Consistency
Does the same information match across different places? If one system lists "NY" and another lists "New York" for the same customer, that's inconsistent. Your data should tell one coherent story, not contradict itself.
Timeliness
Is the data current enough for your needs? Stock prices from yesterday might be worthless today, while historical census data could remain useful for decades. Timeliness depends on context—what you're trying to solve.
Validity
Does the data follow the rules it should? If a "Yes/No" field contains "Maybe" or a phone number has letters, that violates validity. Think of this as data staying within its lane—respecting expected formats and constraints.
Putting It Together
These dimensions often overlap. Data can be complete but inaccurate, or timely but inconsistent. Evaluating all five dimensions helps you spot problems before they derail your analysis.
Key Takeaway: Data quality depends on five dimensions—accuracy, completeness, consistency, timeliness, and validity—that help you judge whether data is trustworthy enough to drive decisions.