Skip to content

a5c-ai/babysitter / library/specializations/data-engineering-analytics/skills/data-quality-profiler

data-quality-profiler

Verified

Profiles data assets to assess quality dimensions, detect anomalies, and generate comprehensive data quality reports with actionable recommendations.

Data Quality Profiler

Profiles data assets to assess quality dimensions and detect anomalies across the six core data quality dimensions.

Overview

This skill performs comprehensive data profiling to assess completeness, accuracy, consistency, validity, timeliness, and uniqueness. It generates statistical profiles, detects anomalies, identifies PII, and provides actionable recommendations for data quality improvement.

Capabilities

  • Statistical profiling - Distributions, cardinality, null percentages, min/max values
  • Data type inference and validation - Detect actual vs declared types
  • Pattern detection - Regex patterns, formats, common structures
  • Anomaly detection - Outliers, drift, unexpected values
  • Referential integrity checking - Foreign key validation
  • Freshness monitoring - Data age and update frequency
  • Volume trend analysis - Record count patterns over time
  • Schema change detection - Structural changes between runs
  • Cross-column correlation analysis - Identify dependent columns
  • PII detection and classification - Sensitive data identification

Input Schema

{
  "dataSource": {
    "type": "object",
    "required": true,
    "properties": {
      "type": {
        "type": "string",
        "enum": ["table", "file", "query"],
        "description": "Type of data source"
      },
      "connection": {
        "type": "object",
        "description": "Connection details (platform, database, schema)"
      },
      "identifier": {
        "type": "string",
        "description": "Table name, file path, or query string"
      }
    }
  },
  "sampleSize": {
    "type": "number",
    "description": "Number of rows to sample (null for full scan)",
    "default": 10000
  },
  "dimensions": {
    "type": "array",
    "items": {
      "type": "string",
      "enum": ["accuracy", "completeness", "consistency", "validity", "timeliness", "uniqueness"]
    },
    "default": ["completeness", "validity", "uniqueness"],
    "description": "Quality dimensions to assess"
  },
  "previousProfile": {
    "type": "object",
    "description": "Previous profile for drift detection"
  },
  "businessRules": {
    "type": "array",
    "items": {
      "column": "string",
      "rule": "string",
      "threshold": "number"
    },
    "description": "Custom business rules to validate"
  },
  "piiDetection": {
    "type": "boolean",
    "default": true,
    "description": "Enable PII detection and classification"
  }
}

Output Schema

{
  "profile": {
    "type": "object",
    "properties": {
      "tableName": "string",
      "rowCount": "number",
      "columnCount": "number",
      "profileTimestamp": "string",
      "columns": {
        "type": "array",
        "items": {
          "name": "string",
          "declaredType": "string",
          "inferredType": "string",
          "statistics": {
            "nullCount": "number",
            "nullPercent": "number",
            "distinctCount": "number",
            "distinctPercent": "number",
            "min": "any",
            "max": "any",
            "mean": "number",
            "median": "number",
            "stddev": "number",
            "histogram": "array"
          },
          "patterns": {
            "mostCommon": "array",
            "detectedFormat": "string",
            "regexPattern": "string"
          },
          "qualityScores": {
            "completeness": "number",
            "validity": "number",
            "uniqueness": "number"
          }
        }
      }
    }
  },
  "anomalies": {
    "type": "array",
    "items": {
      "column": "string",
      "type": "outlier|drift|unexpected_null|unexpected_value|format_violation",
      "severity": "high|medium|low",
      "description": "string",
      "examples": "array",
      "recommendation": "string"
    }
  },
  "piiFindings": {
    "type": "array",
    "items": {
      "column": "string",
      "piiType": "email|phone|ssn|credit_card|name|address|ip|custom",
      "confidence": "number",
      "sampleCount": "number",
      "recommendation": "string"
    }
  },
  "overallScore": {
    "type": "number",
    "description": "Weighted quality score (0-100)"
  },
  "dimensionScores": {
    "completeness": "number",
    "accuracy": "number",
    "consistency": "number",
    "validity": "number",
    "timeliness": "number",
    "uniqueness": "number"
  },
  "recommendations": {
    "type": "array",
    "items": {
      "priority": "high|medium|low",
      "category": "string",
      "description": "string",
      "impact": "string"
    }
  },
  "drift": {
    "type": "object",
    "description": "Changes compared to previous profile",
    "properties": {
      "schemaChanges": "array",
      "statisticalDrift": "array",
      "volumeChange": "object"
    }
  }
}

Usage Examples

Basic Table Profiling

{
  "dataSource": {
    "type": "table",
    "connection": {
      "platform": "snowflake",
      "database": "analytics",
      "schema": "core"
    },
    "identifier": "dim_customers"
  },
  "dimensions": ["completeness", "validity", "uniqueness"]
}

File Profiling with PII Detection

{
  "dataSource": {
    "type": "file",
    "identifier": "./data/customer_export.csv"
  },
  "sampleSize": 50000,
  "piiDetection": true,
  "dimensions": ["completeness", "validity", "accuracy"]
}

Query-Based Profiling with Business Rules

{
  "dataSource": {
    "type": "query",
    "connection": {
      "platform": "bigquery",
      "project": "my-project"
    },
    "identifier": "SELECT * FROM orders WHERE order_date >= '2024-01-01'"
  },
  "businessRules": [
    {"column": "order_total", "rule": "positive", "threshold": 0},
    {"column": "status", "rule": "in_set", "values": ["pending", "completed", "cancelled"]},
    {"column": "customer_id", "rule": "not_null", "threshold": 100}
  ]
}

Drift Detection

{
  "dataSource": {
    "type": "table",
    "identifier": "fact_sales"
  },
  "previousProfile": {
    "profileTimestamp": "2024-01-01T00:00:00Z",
    "rowCount": 1000000,
    "columns": [...]
  },
  "dimensions": ["consistency", "timeliness"]
}

Quality Dimensions

Completeness (0-100)

Measures the presence of required data:

MetricCalculation
Column completeness(total - nulls) / total * 100
Row completenessrows with all required fields / total rows * 100
OverallWeighted average across columns

Validity (0-100)

Measures conformance to business rules:

Check TypeExample
Type conformanceString in INT column
Format conformanceInvalid email format
Range conformanceAge > 150
ReferentialFK without matching PK

Uniqueness (0-100)

Measures duplicate and cardinality:

MetricCalculation
Distinct ratiodistinct / total * 100
Duplicate counttotal - distinct
PK uniquenessunique PKs / total * 100

Accuracy (0-100)

Measures correctness against ground truth:

  • Requires reference data or business rules
  • Cross-validates related columns
  • Checks mathematical relationships

Consistency (0-100)

Measures uniformity across the dataset:

  • Format consistency (dates, numbers)
  • Categorical value consistency
  • Cross-column logical consistency

Timeliness (0-100)

Measures data freshness:

MetricThreshold
Data ageHours since last update
Freshness SLA% meeting freshness requirement
Lag detectionProcessing delay measurement

PII Detection Categories

TypePattern Examples
Emailxxx@domain.com
Phone(XXX) XXX-XXXX, +1-XXX-XXX-XXXX
SSNXXX-XX-XXXX
Credit CardXXXX-XXXX-XXXX-XXXX (with Luhn check)
NameFirst/Last name patterns
AddressStreet, city, state, zip patterns
IP AddressIPv4 and IPv6

Integration Points

MCP Server Integration

  • Elementary MCP - Data observability and anomaly detection
  • Database MCPs - Direct profiling queries

Related Skills

  • Great Expectations Generator (SK-DEA-006) - Generate expectation suites from profiles
  • Data Catalog Enricher (SK-DEA-017) - Enrich catalog with profile metadata

Applicable Processes

  • Data Quality Framework (data-quality-framework.js)
  • Data Catalog (data-catalog.js)
  • ETL/ELT Pipeline (etl-elt-pipeline.js)
  • A/B Testing Pipeline (ab-testing-pipeline.js)

References

Version History

  • 1.0.0 - Initial release with six dimension profiling and PII detection
slack-gif-creatoranthropics/skillsKnowledge and utilities for creating animated GIFs optimized for Slack. Provides constraints, validation tools, and animation concepts. Use when users request animated GIFs for Slack like "make me a GIF of X doing Y for Slack."Stars: 196kUpdated Claude CodeCodexVerifiedInstall
ponytailDietrichGebert/ponytailLazy senior dev mode for any coding task: the smallest change that fully solves it, and a reply a busy human understands in one read.Stars: 193kUpdated Claude CodeCodexVerifiedInstall
ponytail-debtDietrichGebert/ponytailHarvest every shortcut: comment into one debt ledger, so deferrals get tracked instead of forgotten. One-shot report.Stars: 193kUpdated Claude CodeCodexVerifiedInstall
ponytail-debtDietrichGebert/ponytailList every `shortcut:` comment (and older `ponytail:` ones) as a debt ledger. One-shot report, changes nothing. Use for "ponytail debt", "what did ponytail defer", "list the shortcuts", /ponytail-debt.Stars: 193kUpdated Claude CodeCodexVerifiedInstall