1. Enter assessment data
Use passed and failed record counts or percentages for each quality dimension.
Data management tool
Measure data quality across seven critical dimensions
Calculate a transparent weighted score for completeness, accuracy, validity, consistency, uniqueness, timeliness and integrity, then prioritise remediation by threshold gap.
Transparent and deterministic: the same inputs produce the same result. Scores depend on user-supplied information and should be validated against business context.
How it works
From assessment inputs to prioritised action
2. Apply weights
Use equal weighting or assign custom weights totalling exactly 100%.
3. Review priorities
See threshold breaches, weighted contribution and practical remediation actions.
Interactive calculator
Data-quality assessment
Complete all seven dimensions. Basic submission and result generation work without JavaScript.
Methodology and limitations
Use the score as a decision aid, not a substitute for governance
Score scale
A: 95–100, B: 90–94.99, C: 80–89.99, D: 70–79.99, and E: below 70. Dimension thresholds are separately defined and may be stricter than the overall grade bands.
Assessment boundaries
Define the dataset, time period, rules, population and sample before entering results. Comparing scores is meaningful only when those boundaries remain consistent.
Weight selection
Equal weighting is useful for an initial view. Custom weighting is more appropriate when business impact differs, such as giving integrity or accuracy greater weight for regulated reporting.
Known limitations
The calculator does not independently verify data, test rule design, detect sampling bias or assess whether a dataset is fit for a particular decision, model or regulatory obligation.
Privacy note: no external API is used. Basic form submission is processed by the hosting server, and CSV or JSON exports are generated locally in your browser. No user data should be stored unless the existing site deliberately implements secure server-side storage.
Related tools
Continue your data and AI assessment
Frequently asked questions
Practical guidance for using the calculator
What does the overall data-quality score represent?
It is the weighted average of the seven dimension scores. It provides a concise portfolio view while preserving dimension-level detail for remediation.
Can I use passed and failed record counts?
Yes. In count mode, passed plus failed records must equal the total records assessed for every dimension.
Can I enter percentages instead of counts?
Yes. Percentage mode accepts a value from 0 to 100 for each dimension and is useful when scores already exist.
How are custom weights handled?
Every custom weight must be between 0 and 100, and the seven weights must total 100% within a 0.01 percentage-point tolerance.
What score thresholds are used?
Overall grades are A at 95 or above, B from 90 to 94.99, C from 80 to 89.99, D from 70 to 79.99 and E below 70.
Why do dimension thresholds differ?
The defaults reflect common operational expectations. Your organisation may need stricter or more flexible thresholds based on risk, regulation and use-case criticality.
Does a high score guarantee fitness for use?
No. Fitness also depends on lineage, coverage, rule quality, sampling, context and the consequences of an error.
How should I select the assessed records?
Use the full population where feasible. Otherwise use a representative, documented sample that covers relevant sources, time periods, segments and exception cases.
How often should data quality be measured?
Set the cadence according to data volatility and business impact, from continuous monitoring for operational feeds to periodic checks for stable reference data.
What should I do after a threshold breach?
Prioritise the largest gaps tied to critical processes, identify root causes at source, assign accountable owners and verify improvements through repeat measurement.
Is information sent to external services?
No external APIs or third-party libraries are used. Exports are produced locally in the browser.
Can I compare scores over time?
Yes, provided the data scope, rules, thresholds, weighting and sampling approach remain consistent and material changes are documented.