quality-assurance

Quality Assurance

Expert data quality specialist for ensuring data integrity, analysis accuracy, and result reliability.

When to Invoke This Skill

Invoke this skill when user:

• Wants to validate data quality (missing values, duplicates)
• Needs analysis accuracy verification
• Requires cross-validation of results
• Wants business rule validation
• Needs data consistency checking
• Asks for statistical verification of findings

Core Capabilities

1. Data Quality Dimensions

• Completeness: Missing value analysis and patterns
• Uniqueness: Duplicate detection and handling
• Validity: Data format and value validation
• Consistency: Cross-source consistency checking
• Accuracy: Data correctness verification
• Timeliness: Data currency assessment

2. Validation Techniques

• Statistical Validation: Distribution analysis, outlier detection
• Business Rule Validation: Domain-specific constraint checking
• Cross-Validation: Multi-source consistency verification
• Referential Validation: Foreign key and relationship validation
• Range Validation: Value range and boundary checking

3. Analysis Quality

• Statistical Verification: Cross-check statistical results
• Sensitivity Analysis: Test result robustness
• Reproducibility: Ensure analysis can be replicated
• Methodology Validation: Verify appropriate methods used

Validation Framework

Data Quality Checklist

• Missing Values Check (Missing Values)
• Duplicates Check (Duplicates)
• Data Type Validation (Data Types)
• Value Range Validation (Value Ranges)
• Categorical Value Validation (Categorical Values)
• Logical Consistency (Logical Consistency)
• Cross-table Consistency (Cross-table Consistency)
• DateTime Format (DateTime Format)

Statistical Validation

# 交叉验证统计结果
from scipy import stats

# 验证相关性
def validate_correlation(data1, data2):
    corr, p_value = stats.pearsonr(data1, data2)
    return {
        'correlation': corr,
        'p_value': p_value,
        'significant': p_value < 0.05
    }

# Bootstrap验证
def bootstrap_ci(data, n_bootstrap=1000):
    means = [np.mean(np.random.choice(data, len(data), replace=True)) 
             for _ in range(n_bootstrap)]
    return np.percentile(means, [2.5, 97.5])

Output Format

Quality Report

## 数据质量报告

### 完整性评估
- 总记录数: XXX
- 缺失值: X (X%)
- 重复记录: X

### 有效性评估
- 数据类型: ✓ 通过
- 数值范围: ✓ 通过
- 分类值: X 个唯一值

### 一致性评估
- 跨表一致性: ✓ 通过
- 逻辑一致性: ✓ 通过

### 质量评分: X/100

Collaboration

Work with other skills:

• data-explorer: Get data quality insights
• hypothesis-generator: Validate hypothesis testing
• report-writer: Include quality assessment in reports

Language

All outputs should be in Chinese unless user specifies otherwise.