data-validate

name: data-validate description: Validate data and analysis before sharing - methodology, accuracy, bias, and data quality checks user-invocable: true argument-hint: "" allowed-tools: [mcp__qsv__qsv_sniff, mcp__qsv__qsv_count, mcp__qsv__qsv_headers, mcp__qsv__qsv_index, mcp__qsv__qsv_stats, mcp__qsv__qsv_moarstats, mcp__qsv__qsv_frequency, mcp__qsv__qsv_search, mcp__qsv__qsv_select, mcp__qsv__qsv_slice, mcp__qsv__qsv_sqlp, mcp__qsv__qsv_joinp, mcp__qsv__qsv_command, mcp__qsv__qsv_list_files, mcp__qsv__qsv_search_tools, mcp__qsv__qsv_get_working_dir, mcp__qsv__qsv_set_working_dir]

Data Validate

Validate data files and analyses for accuracy, methodology, and potential biases before sharing with stakeholders. Generates a confidence assessment and improvement suggestions.

Cowork note: If relative paths don't resolve, call mcp__qsv__qsv_get_working_dir and mcp__qsv__qsv_set_working_dir to sync the working directory.

Usage

The input can be:

• A CSV/TSV/Excel file to validate for data quality
• An analysis, report, or document to review for methodology
• SQL queries and their results to verify
• A description of methodology and findings

Steps

1. Data Quality Validation

If a data file is provided, run these checks using qsv:

a. Index and profile: Run mcp__qsv__qsv_index, then mcp__qsv__qsv_stats with cardinality: true, stats_jsonl: true and mcp__qsv__qsv_sniff to understand the data.

b. Completeness: Read .stats.csv — check nullcount and sparsity for each column. Flag columns with sparsity > 0.5.

c. Uniqueness: Compare cardinality to row count from mcp__qsv__qsv_count. Flag key columns (ID, email) where cardinality < row count. Run mcp__qsv__qsv_command with command: "dedup" and options: {"dupes-output": "dupes.csv"} to find exact duplicates.

d. Validity: Check type column in stats — flag String columns that should be numeric. Run mcp__qsv__qsv_command with command: "validate" against a JSON Schema if available.

e. Consistency: Run mcp__qsv__qsv_frequency with limit: 20 on categorical columns — look for case variants ("NYC" vs "nyc"), inconsistent formats, unexpected values.

f. Accuracy: Read .stats.csv for min, max, mean, stddev — flag implausible ranges (negative ages, latitude > 90, future dates). Run mcp__qsv__qsv_moarstats with advanced: true — check outliers_percentage > 5%, kurtosis > 10 (extreme outliers).

g. Distribution sanity: Read moarstats columns for deeper validation:

• median_mean_ratio — if < 0.8 or > 1.2, distribution is significantly skewed; verify the mean isn't misleading
• winsorized_mean_25pct vs mean — large divergence (> 10%) confirms outliers are distorting the average
• mad (median absolute deviation) — more robust than stddev for outlier detection; if mad_stddev_ratio > 0.8, stddev is reasonably reliable
• jarque_bera_pvalue — if < 0.05, data is NOT normally distributed; flag any analysis that assumes normality
• mode_count — if mode accounts for > 50% of values, investigate whether this reflects a data entry default or missing value masking

h. Join integrity (if multiple files): Run mcp__qsv__qsv_joinp with left_anti: true to find orphaned foreign keys.

i. Injection screening: Run mcp__qsv__qsv_command with command: "searchset", regexset-file: "${CLAUDE_PLUGIN_ROOT}/resources/injection-regexes.txt", and flag: "injection_match" to scan for malicious payloads.

2. Review Methodology and Assumptions

Examine the analysis for:

• Question framing: Is the analysis answering the right question? Could it be interpreted differently?
• Data selection: Are the right datasets being used? Is the time range appropriate?
• Population definition: Is the analysis population correctly defined? Are there unintended exclusions?
• Metric definitions: Are metrics defined clearly and consistently? Do they match how stakeholders understand them?
• Baseline and comparison: Is the comparison fair? Are time periods, cohort sizes, and contexts comparable?

3. Check for Common Analytical Pitfalls

Systematically review against these pitfalls:

4. Verify Calculations and Aggregations

Spot-check using mcp__qsv__qsv_sqlp:

• Recalculate key numbers independently
• Verify subtotals sum to totals: SELECT SUM(subtotal) as check_total FROM data
• Check percentages sum to ~100%: SELECT SUM(pct) FROM data
• Validate YoY/MoM comparisons use correct base periods
• Confirm filters are applied consistently across all metrics

Magnitude Checks

Red Flags That Warrant Investigation

• Any metric that changed by more than 50% period-over-period without an obvious cause
• Counts or sums that are exact round numbers (suggests a filter or default value issue)
• Rates exactly at 0% or 100% (may indicate incomplete data)
• Results that perfectly confirm the hypothesis (reality is usually messier)
• Identical values across time periods or segments (suggests the query is ignoring a dimension)

5. Assess Visualizations (if present)

If the analysis includes charts:

• Do axes start at appropriate values (zero for bar charts)?
• Are scales consistent across comparison charts?
• Do chart titles accurately describe what's shown?
• Could the visualization mislead a quick reader?
• Are there truncated axes, inconsistent intervals, or 3D effects that distort perception?

6. Evaluate Narrative and Conclusions

Review whether:

• Conclusions are supported by the data shown
• Alternative explanations are acknowledged
• Uncertainty is communicated appropriately
• Recommendations follow logically from findings
• The level of confidence matches the strength of evidence

7. Suggest Improvements

Provide specific, actionable suggestions:

• Additional analyses that would strengthen the conclusions
• Caveats or limitations that should be noted
• Better visualizations or framings for key points
• Missing context that stakeholders would want

8. Generate Confidence Assessment

Rate the analysis on a 3-level scale:

Ready to share — Analysis is methodologically sound, calculations verified, caveats noted. Minor suggestions for improvement but nothing blocking.

Share with noted caveats — Analysis is largely correct but has specific limitations or assumptions that must be communicated to stakeholders. List the required caveats.

Needs revision — Found specific errors, methodological issues, or missing analyses that should be addressed before sharing. List the required changes with priority order.

Pre-Delivery QA Checklist

Data Quality Checks

• Source verification: Confirmed data sources. Run mcp__qsv__qsv_sniff to verify format and encoding.
• Freshness: Data is current enough. Noted the "as of" date.
• Completeness: No gaps in time series. Check nullcount/sparsity in .stats.csv.
• Null handling: Nulls handled appropriately (excluded, imputed, or flagged).
• Deduplication: No double-counting. Verify with mcp__qsv__qsv_count before/after joins, dedup --dupes-output.
• Filter verification: All filters correct. No unintended exclusions.

Calculation Checks

• Aggregation logic: GROUP BY includes all non-aggregated columns.
• Denominator correctness: Rate calculations use the right denominator (non-zero).
• Date alignment: Comparisons use same time period length. Partial periods excluded or noted.
• Join correctness: JOIN types appropriate. Verify row counts with mcp__qsv__qsv_count after joins.
• Metric definitions: Metrics match stakeholder definitions. Deviations noted.
• Subtotals sum: Parts add up to the whole. Verify with mcp__qsv__qsv_sqlp.

Reasonableness Checks

• Magnitude: Numbers in plausible range. Check min/max in .stats.csv.
• Trend continuity: No unexplained jumps. Use mcp__qsv__qsv_sqlp to check period-over-period.
• Cross-reference: Key numbers match other known sources.
• Edge cases: Checked boundaries — empty segments, zero-activity periods, new entities.

Presentation Checks

• Chart accuracy: Bar charts start at zero. Axes labeled. Scales consistent.
• Number formatting: Appropriate precision and consistent formatting.
• Title clarity: Titles state the insight, not just the metric. Date ranges specified.
• Caveat transparency: Known limitations and assumptions stated explicitly.
• Reproducibility: Someone else could recreate this analysis.

Common Analytical Pitfalls (Reference)

Join Explosion

A many-to-many join silently multiplies rows, inflating counts and sums. Detect: mcp__qsv__qsv_count before and after join — if count increased, investigate the join relationship. Prevent: Use COUNT(DISTINCT id) instead of COUNT(*) when counting entities through joins.

Survivorship Bias

Analyzing only entities that exist today, ignoring churned/deleted/failed ones. Detect: mcp__qsv__qsv_frequency on status columns — are all lifecycle states represented? Prevent: Ask "who is NOT in this dataset?" before drawing conclusions.

Incomplete Period Comparison

Comparing a partial period to a full period. Detect: mcp__qsv__qsv_sqlp to check min/max dates per period. Prevent: Filter to complete periods or compare same number of days.

Denominator Shifting

The denominator changes between periods, making rates incomparable. Detect: mcp__qsv__qsv_sqlp to verify denominator definition consistency. Prevent: Use consistent definitions across all compared periods.

Average of Averages

Averaging pre-computed averages gives wrong results when group sizes differ. Detect: Compare mcp__qsv__qsv_stats mean against mcp__qsv__qsv_sqlp weighted average. Prevent: Always aggregate from raw data.

Simpson's Paradox

Trend reverses when data is aggregated vs. segmented. Detect: mcp__qsv__qsv_sqlp GROUP BY at different granularity levels — does the conclusion change? Prevent: Always check results at segment level before aggregating.

Report Format

## Validation Report

### Overall Assessment: [Ready to share | Share with caveats | Needs revision]

### Data Quality Summary
- File: [format, rows, columns, encoding]
- Completeness: [null rates, gaps found]
- Uniqueness: [duplicates found, cardinality issues]
- Validity: [type mismatches, schema violations]
- Accuracy: [outliers, implausible ranges]

### Methodology Review
[Findings about approach, data selection, definitions]

### Issues Found
1. [Severity: High/Medium/Low] [Issue description and impact]
2. ...

### Calculation Spot-Checks
- [Metric]: [Verified / Discrepancy found]
- ...

### Visualization Review
[Any issues with charts or visual presentation]

### Suggested Improvements
1. [Improvement and why it matters]
2. ...

### Required Caveats for Stakeholders
- [Caveat that must be communicated]
- ...

Notes

• Run this command before any high-stakes presentation or decision
• For pure data quality profiling without methodology review, use /data-profile instead
• If validation finds data quality issues, use /data-clean to fix them and re-validate
• Share the validation report alongside your analysis to build stakeholder confidence