> AI agents: this is one page from Mammoth Analytics documentation. The index of all pages as Markdown is https://docs.mammoth.io/llms.txt. Append `.md` to any docs URL, or send `Accept: text/markdown`, to get Markdown. # Data Quality Report > The Data Quality Report scores a View against six DAMA dimensions, lists critical issues and warnings, and recommends pipeline tasks to fix them. The Data Quality Report evaluates the health of a View's data using the DAMA (Data Management Association) framework. It analyzes every column and gives an overall quality score, critical issues and warnings in plain language, and recommended pipeline tasks. The report doesn't change your data; you build the recommended tasks yourself. ![Data Quality dialog Overview section with a left navigation listing Overview, Critical Issues, Quality Warnings, Recommended Actions and Automation, an 87% quality score, and tiles for Completeness, Validity, Uniqueness, Consistency, Accuracy and Timeliness](https://docs.mammoth.io/api/v1/images/20261009_195501_DataQualityOverview.jpg) The Data Quality Report complements Explore Cards: Explore Cards show patterns in individual columns, and the report diagnoses the whole View. ## When to generate quality reports **Before building a pipeline** Generate a quality report immediately after uploading data to identify show-stopper issues before investing time in transformations. If your data scores "Poor" (below 60%), focus on critical issues first. **Before production** Run a quality assessment before sharing data with stakeholders or feeding into dashboards. A score of "Good" (80%+) typically indicates the data is ready for production use. **Compliance documentation** Generate quality reports for regulatory requirements, audit trails, and data governance documentation. The DAMA framework provides industry-standard quality metrics. **Migration quality assurance** When migrating from legacy systems, track quality score improvement throughout the process. Set acceptance criteria (e.g., minimum 90% score) before go-live. ## Generate a Data Quality Report Prerequisites: the pipeline is not running, and you are not in preview or edit mode. The button is disabled in those cases and in View-only workspaces. The button appears only when Data Quality is enabled for your workspace. 1. Open the View. 2. Click **View Data Quality** in the footer of the View. Once a report exists, the button reads **Data Quality Score** followed by the score level (for example, *Good*). Generating a report opens the summary popover when the analysis finishes. If the analysis fails, Mammoth shows an error message and the button stays as it was. Mammoth analyzes all columns in the View automatically, so you don't select columns. ![Data grid with a Data Quality popup at the bottom showing score 84/100, 1 critical issue and a View Details button](https://docs.mammoth.io/api/v1/images/20260418_181107_ViewtoolbarshowingDataQuality.jpg) ### What happens during analysis Mammoth builds column statistics for the View's data (completeness, uniqueness, minimum and maximum, sample values, and so on). AI then classifies each column by its likely role (identifier, measure, dimension, date, or text) and flags likely problems based on what each column is for. Mammoth checks those flags against the actual statistics and removes any the data doesn't support. While the analysis runs, the button reads "Analyzing Data Quality" with a progress indicator. The report evaluates the View's data after its pipeline, not the Explore Cards filters you have applied. ## Report interface Click the button to see the summary popover, titled **Data Quality** with the score level. It shows the **Score** out of 100, **Column completeness**, the number of **Critical Issues**, and **Recommendations** (as a task count). Click **View Details** to open the full report, with sections in the left navigation. Which sections appear depends on what the analysis finds. A View with no detected issues shows only the Overview: - **Overview**—Overall score and dimension-by-dimension summary (always shown) - **Critical Issues**—High-severity problems requiring immediate attention (shown only when critical issues are found) - **Quality Warnings**—Medium-severity issues worth addressing (shown only when warnings are found) - **Recommended Actions**—Prioritized pipeline tasks to resolve identified issues (shown only when there are recommendations) - **Automation**—Workflow suggestions for systematic remediation (shown only when the analysis has automation suggestions); see [Automation section](#automation-section) ![Data Quality dialog Overview section with a left navigation listing Overview, Critical Issues, Quality Warnings, Recommended Actions and Automation, an 87% quality score, and tiles for Completeness, Validity, Uniqueness, Consistency, Accuracy and Timeliness](https://docs.mammoth.io/api/v1/images/20261009_195501_DataQualityOverview.jpg) ## Overview section The Overview displays your overall quality score and its status, and the six DAMA dimensions as individual score tiles. Score display example from the store_transaction dataset (669,430 rows, 14 columns): ```plaintext Overall Score: 44% — Poor (Critical remediation required) ``` `Completeness: 100% All columns complete`\ `Validity: 28% 10 validity issues`\ `Uniqueness:—1 identifier(s) with duplicates`\ `Consistency: 66% 2 case inconsistencies`\ `Accuracy: 60% 4 columns with outliers`\ `Timeliness: 25% Data is 1678 days old (very stale vs expected 1 day)` Note: A dimension that scores 0% displays as "-" (dash) rather than a number. For Uniqueness this happens when every identifier column has duplicates, or when the View has no identifier columns. The note below the dash explains which. ## The DAMA framework The Data Management Association (DAMA) framework provides six industry-standard quality dimensions. Mammoth scores each dimension from 0 to 100% and combines them into an overall quality score. Dimension scores are based on how many columns have a problem, not on how many rows. ### 1. Completeness—missing data assessment **What it measures:** Null values, missing data, empty fields **Score**: The percentage of columns with no missing values. **Typical issues**: - Required columns (such as keys) that contain empty values - Columns that are entirely empty **Business impact examples:** - Cannot join customer data if Customer_ID has nulls - Revenue calculations fail with missing transaction amounts - Reporting gaps in time series with missing dates ### 2. Uniqueness—duplicate detection **What it measures:** Duplicate values in columns that should be unique **Score**: The percentage of identifier columns (IDs and keys) without duplicates. If the View has no identifier columns, the score is 0% and shows as a dash. **Typical issues**: - Duplicates in an identifier column, flagged when it is less than 95% unique **Real-world example:** ```plaintext Critical: Transaction ID has low uniqueness (4.11% unique) — approx. 641,912 duplicate rows detected Remove duplicate rows by key to ensure reliable joins ``` **Business impact:** Double-counting in metrics, inflated KPIs, analysis errors, unreliable joins ### 3. Timeliness—data freshness **What it measures:** Data age relative to expected refresh frequency **Score**: Compares the age of the newest date in your date columns against the refresh frequency the AI expects for that kind of data (for example, daily for transactions). The score is 100% when the data is within the expected age, 70% when it is up to twice that age, 50% up to four times, and 25% beyond that. A View with no date columns scores 100%. If the AI gives no expected cadence for the date columns, the score is 100% when the data is up to 30 days old, 70% up to 90 days, and 50% beyond that. The report states the measured age and, when there is one, the expected cadence in plain language. **Example**: ```plaintext Timeliness: 25% "Data is 1678 days old (very stale vs expected 1 day)" ``` **Common problems:** - ETL jobs failing silently (data stops updating) - Source systems that stopped sending recent data ### 4. Validity—unexpected values **What it measures:** Values that are unexpected for what the column holds **Score**: The percentage of columns without unexpected negatives, unexpected zeros, or future dates. **Typical issues**: - Negative values in a column that should not be negative (prices, quantities) - Zero values in a column that should not contain zero - Future dates The report does not check text formats such as email or phone patterns. ### 5. Accuracy—outlier detection **What it measures:** Numeric values that fall outside a reasonable range **Score**: The percentage of numeric columns without outliers. Outliers are identified with the interquartile range (IQR) method. This score never drops below 60%, because outliers need domain knowledge to judge. A View with no numeric columns scores 100%. **Example issue output:** ```plaintext Warning: Price has potential outliers (1 detected). Investigate and validate outliers to protect accuracy of price analytics. ``` ### 6. Consistency—casing uniformity **What it measures:** Casing uniformity in text columns **Score**: The percentage of text columns without case variants (for example, "Active", "ACTIVE", and "active"). A View with no text columns scores 100%. The report does not check relationships between columns. **Consistency examples:** ```plaintext Warning: Department has inconsistent casing across values — standardize text case to enable reliable grouping and reporting Warning: Category has inconsistent casing across values — standardize text case to enable reliable segmentation ``` ## Quality scoring system The overall score is a weighted composite across all six dimensions: - Validity: 25% weight - Uniqueness: 25% weight - Completeness: 20% weight - Consistency: 15% weight - Accuracy: 10% weight - Timeliness: 5% weight The overall score is the weighted average, rounded down to a whole number. ### Score tiers with recommended actions **Excellent (95-100%)** - Status: Production-ready - Action: Proceed with confidence - Use cases: Dashboard publishing, stakeholder sharing, automated reporting - Typical issues: Minor formatting inconsistencies only **Good (80-94%)** - Status: Acceptable with awareness - Action: Note limitations, consider minor cleanup - Use cases: Internal analysis, exploratory work, prototype dashboards - Typical issues: Some missing values, formatting inconsistencies, minor duplicates **Fair (60-79%)** - Status: Significant issues requiring attention - Action: Address critical issues before production use - Use cases: Development work only, not recommended for stakeholder sharing - Typical issues: Multiple data quality dimensions affected, some critical issues **Poor (<60%)** - Status: Major problems, extensive cleanup required - Action: Systematic remediation needed - Use cases: Quality assessment only, not suitable for analysis - Typical issues: Multiple critical issues, significant business impact ![Data Quality popup reading Poor, score 44/100, 15 critical issues, 4 recommended tasks and a View Details button](https://docs.mammoth.io/api/v1/images/20260418_181338_DataQualityScoreindicatorinbottomstatus1.jpg) ## Critical issues section The Critical Issues section lists all High Severity problems detected across your dataset. Each issue card includes the issue name, a plain-language description of what was found, the row counts or percentages affected, and the business impact of leaving it unresolved. Example Critical Issues from the store_transaction dataset: ```plaintext Transaction ID low uniqueness — High Severity Transaction ID has low uniqueness (4.11% unique). Remove duplicate rows by key to ensure reliable joins. ``` `Price unexpected negatives—High Severity`\ `Price contains unexpected negative values (1,210 negative). Review and`\ `correct negative prices to ensure valid monetary reporting.` `Price unexpected zeros—High Severity`\ `Price contains unexpected zero values (7,400 zero). Review and correct`\ `zero prices to avoid distorted averages and revenue.` `Subtotal unexpected negatives—High Severity`\ `Subtotal contains unexpected negatives (1,230 negative). Review negative`\ `subtotals to prevent incorrect revenue and margin calculations.` The report first shows as many critical issues as fit in the window. Click **Load more** to see the rest. ![Data Quality dialog Critical Issues section listing three High Severity issue cards: order_date stale data, quantity missing values and sales missing values](https://docs.mammoth.io/api/v1/images/20261009_195515_DataQualityCritical.jpg) ## Quality warnings section Quality Warnings lists Medium severity issues that do not block analysis but can degrade report accuracy or data reliability if left unaddressed. Example Quality Warnings from the store_transaction dataset: ```plaintext Time stale data — Medium Time contains stale entries (1678 days old median). Filter or refresh time values to restore recency for time-based analysis. ``` `Department case inconsistency—Medium`\ `Department has inconsistent casing across values. Standardize text case`\ `to enable reliable grouping and reporting.` `Category case inconsistency—Medium`\ `Category has inconsistent casing across values. Standardize text case`\ `to enable reliable segmentation.` `Price potential outliers—Medium`\ `Price has potential outliers (1 detected). Investigate and validate`\ `outliers to protect accuracy of price analytics.` The report first shows as many warnings as fit in the window. Click **Load more** to see the rest. ![Data Quality dialog Quality Warnings section listing one Medium severity card: city missing values](https://docs.mammoth.io/api/v1/images/20261009_195541_DataQualityWarnings.jpg) ## Recommended actions section The Recommended Actions section translates identified issues into specific pipeline tasks (typically 3 to 5), each with a priority. Each action card describes what the task will do and which issues it addresses. Suggested tasks come from tasks such as Remove Duplicate Rows, Conditional Filter, Text Formatting, Bulk Replace, Find & Replace, Fill Missing Values, Remove Column, and Math Function. Action cards are read-only descriptions. Use them as guidance to build the corresponding pipeline tasks yourself. Example Recommended Actions from the store_transaction dataset: ```plaintext Remove Duplicate Rows — High Priority Remove duplicate rows where Transaction ID repeats (approx. 641,912 duplicate rows) and keep one canonical record per transaction to ensure reliable joins. ``` `Conditional Filter—High Priority`\ `Filter rows with negative monetary amounts (Price 1,210; Subtotal 1,230;`\ `Tax 1,230; Total 1,470) and investigate before inclusion to prevent`\ `incorrect financial aggregates.` `Math Function—High Priority`\ `Recompute Total = Subtotal + Tax - Discount and flag mismatches`\ `(33,471 subtotal/tax/total outliers) to correct calculation errors for`\ `accurate reporting.` `Text Formatting—Medium Priority`\ `Apply text formatting to Department and Category to normalize casing`\ `across values so grouping and segmentation work reliably.` ## Automation section The Automation section appears in the left navigation (for example, "Automation (2)") only when the report includes workflow suggestions, which are generated only when high-severity issues exist. Each suggestion groups related remediation steps into a single named workflow, shown as a card with a title and description. Note: each automation card is guidance only—there is currently no **Configure** button to act on it directly from the card. Use the description as guidance to manually build the corresponding pipeline tasks. Example Automation suggestion from the store_transaction dataset: ```plaintext Monetary Validation & Repair Automate filtering of negative/zero monetary rows, recalculate totals with Math, and flag mismatches for review to ensure accurate financial reporting. ``` ## Refresh behavior The Data Quality Report is generated when you click the button. When the View's data changes, the saved report is cleared and the button returns to **View Data Quality**. Click it to generate a new report. When to regenerate: - After applying pipeline changes to measure improvement - When new data batches are added to the Dataset - Before final production deployment - For compliance documentation at project milestones A report is saved with the View and reflects a snapshot at generation time. A new report replaces the previous one. ## Recommended workflow **Stage 1—Initial assessment** Generate the report immediately after data upload. Understand baseline quality before any transformations. Identify show-stopper issues early. **Stage 2—Mid-pipeline validation** Regenerate after major transformations. Validate improvements. Identify remaining issues. **Stage 3—Pre-production quality gate** Final report before sharing with stakeholders. Verify data meets business quality standards. Required for regulated industries. ## Real-world use cases ### Pre-production data validation Scenario: data engineer preparing monthly sales report for executives Workflow: 1. Load transformed sales data into Mammoth View 2. Generate Data Quality Report 3. Result: Score 72% (Fair) with critical issues—duplicate orders, negative price values, missing sales rep data 4. Remediation: Build pipeline tasks based on Recommended Actions guidance 5. Validation: Regenerate report to confirm score improvement 6. Outcome: Published dashboard with confidence, avoided an embarrassing executive meeting ### Regulatory compliance documentation Scenario: compliance officer validating quarterly financial reporting data for SOX audit Workflow: 1. Import quarterly financial data from multiple systems 2. Generate Quality Report for audit documentation 3. Assessment: Score 84% (Good)—Completeness 92%, Accuracy 89%, Consistency 78% 4. Documentation: The report's score and dimension breakdown serve as DAMA quality evidence 5. Remediation: Address consistency issues for clean audit trail 6. Outcome: Audit documentation complete ### Data migration quality assurance Scenario: operations team migrating customer data from legacy CRM Workflow: 1. Baseline: Import legacy data—Quality Report shows 43% (Poor) 2. Phase 1: Address critical issues via pipeline transformations—Score improves to 68% (Fair) 3. Phase 2: Address warnings and consistency—Score improves to 87% (Good) 4. Acceptance Testing: Final validation achieves 92% (Excellent) 5. Go-Live: Migration approved based on documented quality improvement Result: quality score improved from 43% to 92%, with a remediation audit trail. ## Data Quality Report versus View Insights ### Data Quality Report - Purpose: Data health diagnosis - Focus: Technical data issues and compliance - Output: Structured problem list with severity levels - Action: Prioritized remediation guidance - Audience: Data engineers, QA teams, compliance officers - When to use: Pre-production validation, compliance docs, troubleshooting ### View Insights (AI-generated) - Purpose: Business intelligence discovery - Focus: Patterns, trends, interesting observations - Output: Natural language business insights - Action: Informational—guides analysis direction - Audience: Business stakeholders, analysts, decision makers - When to use: Business discovery, stakeholder communication, trend identification Run the Quality Report first to ensure data health, then generate Key Insights for business discovery on validated data. ## Advanced quality assessment ### Filtering impact on quality analysis Quality Reports analyze the View's data after its pipeline. Filters you apply in Explore Cards do not change the report. To assess a subset, add a Conditional Filter task to the pipeline, then generate the report again. ### Regeneration strategy When to regenerate: - After applying multiple pipeline fixes to track improvement - When new data batches are added - Before final production deployment - For compliance documentation at project milestones Best practice: regenerate after every 3-4 pipeline tasks applied to track progress and avoid over-fixing. ## Performance and limitations ### Current limitations - Real-time: snapshot analysis only, not continuous monitoring - Cross-dataset: single View analysis only, cannot compare quality across datasets - Recommended Actions: the cards describe tasks but do not create them. Build the tasks in the pipeline - Automation: the Automation section shows workflow suggestions as read-only cards—you cannot act on a suggestion directly from the card ## Troubleshooting quality reports ### Report generation fails Problem: the report does not generate or an error message appears Solutions: - Verify View contains data (not empty) - Wait for any running pipeline to finish, then try again - Try refreshing the page and regenerating ### Unexpected quality scores Problem: score seems too low or too high for the data condition Solutions: - Review individual issue details rather than just the overall score - Check the dimension scores. They count affected columns, so a View with few columns can swing sharply - Check whether the dimension weights match your priorities - Consider business context—some flagged issues may be acceptable for your use case ### A dimension shows a dash (-) What it means: the dimension scored 0%. For Uniqueness, this means every identifier column has duplicates, or the View has no identifier columns, as the note under the dash says. Action: review the Critical Issues section for details on which columns are affected and the extent of duplication. ## Next steps After generating your Data Quality Report: If the score is Excellent (95%+): - Proceed with dashboard creation or stakeholder sharing - Document quality score for audit/compliance records - Set up an Automation to refresh data on a schedule If the score is Fair to Good (60-94%): - Prioritize critical issues using guidance from Recommended Actions - Manually build pipeline tasks corresponding to each recommended action - Regenerate report to measure improvement - Consider acceptable quality threshold for your specific use case If the score is Poor (<60%): - Focus on critical issues first - Use Recommended Actions as your guide for fixes - Do not proceed to analysis until quality improves - Consider addressing data quality at the source system level --- Source: https://docs.mammoth.io/learn/explore-discover/data-quality-report.md · Updated: 2026-10-13