> ## Documentation Index > Fetch the complete documentation index at: https://docs.omophub.com/llms.txt > Use this file to discover all available pages before exploring further. # Get Mapping Quality > Analyze OMOP mapping quality between two vocabularies to assess reliability and trustworthiness of cross-vocabulary translations for production use. This endpoint provides analysis of mapping quality between a specific vocabulary pair, essential for understanding the reliability of cross-vocabulary translations, identifying high-confidence mappings, and assessing the trustworthiness of mapping relationships for clinical and operational use. ## Path Parameters The source vocabulary identifier
**Example:** `SNOMED`, `ICD10CM`, `RxNorm` The target vocabulary identifier
**Example:** `ICD10CM`, `HCPCS`, `NDC` ## Query Parameters Specific vocabulary release version to query
**Example:** `2025.1` ## Response High-level summary of mapping quality Total number of mappings included in analysis Mappings that have quality metadata Overall average confidence score across all mappings Overall quality assessment
**Values:** `excellent`, `good`, `moderate`, `poor` Percentage of mappings with confidence > 0.8 (0-100%) Percentage of officially validated mappings (0-100%) Percentage of disputed or problematic mappings Quality metrics for each vocabulary pair Quality metrics for each source vocabulary Human-readable vocabulary name Quality metrics by target vocabulary Quality metrics for this vocabulary pair Target vocabulary name Number of mappings between these vocabularies Average confidence score for this pair Median confidence score Standard deviation of confidence scores Distribution of equivalence types Rate of validated mappings (0.0-1.0 fraction) Quality rating for this vocabulary pair Main use cases for these mappings Average quality of outgoing mappings Quality analysis by medical domain Quality metrics for each domain Human-readable domain name Number of mappings in this domain Average confidence score for domain Distribution of quality levels Mappings with confidence score of 0.9 or higher Mappings with confidence score between 0.7 and 0.9 Mappings with confidence score between 0.5 and 0.7 Mappings with confidence score below 0.5 Vocabulary pairs with highest quality in this domain Vocabulary pairs with lowest quality in this domain Common quality issues in this domain Detailed quality metrics and distributions Distribution of confidence scores Confidence score ranges and counts Key percentile values Skewness of confidence distribution Kurtosis of confidence distribution Analysis of semantic similarity scores Average semantic similarity score Correlation between similarity and confidence Mappings with low similarity but high confidence Mappings with high similarity but low confidence Quality breakdown by mapping source Quality metrics for official mappings Quality metrics for community mappings Quality metrics for algorithmic mappings Quality metrics for manually created mappings Impact of validation on quality perception Quality comparison between validated and unvalidated Accuracy of validation process Common characteristics of disputed mappings Analysis of quality outliers and anomalies (when include\_outliers=true) Mappings with unusually low quality scores Source concept identifier Target concept identifier Confidence score Why this is considered an outlier Potential quality issues identified Mappings with exceptionally high quality Mappings with inconsistent quality indicators Common patterns in outlier mappings Historical quality trends (when include\_trends=true) Time series of quality improvements Date of measurement Average confidence at this time Validation rate at this time Quality of newly added mappings Annual rate of quality improvement (0.0-1.0 fraction) Areas with most rapid quality improvement Areas with declining quality Recommendations for quality improvement (when include\_recommendations=true) Type of quality improvement recommendation Brief title of the recommendation Detailed description Implementation priority Expected quality improvement Vocabulary pairs that would benefit Estimated implementation effort Analysis metadata and processing information Date when quality analysis was performed Timestamp of underlying mapping data Scope of the quality analysis Time taken to perform analysis Method used to calculate confidence scores ```bash cURL theme={null} curl -X GET "https://api.omophub.com/v1/mappings/quality/SNOMED/ICD10CM" \ -H "Authorization: Bearer YOUR_API_KEY" ``` ```bash cURL (with vocab_release) theme={null} curl -X GET "https://api.omophub.com/v1/mappings/quality/SNOMED/ICD10CM?vocab_release=2025.1" \ -H "Authorization: Bearer YOUR_API_KEY" ``` ```javascript JavaScript theme={null} const response = await fetch('https://api.omophub.com/v1/mappings/quality/SNOMED/ICD10CM', { method: 'GET', headers: { 'Authorization': 'Bearer YOUR_API_KEY' } }); const qualityData = await response.json(); const { quality_overview } = qualityData.data; console.log('=== Mapping Quality Analysis ==='); console.log(`Total mappings analyzed: ${quality_overview.total_mappings_analyzed.toLocaleString()}`); console.log(`Overall quality rating: ${quality_overview.overall_quality_rating}`); console.log(`Average confidence: ${(quality_overview.overall_average_confidence * 100).toFixed(1)}%`); console.log(`High-quality mappings: ${quality_overview.high_quality_percentage.toFixed(1)}%`); console.log(`Validated mappings: ${quality_overview.validated_percentage.toFixed(1)}%`); // Display domain quality console.log('\n=== Quality by Domain ==='); Object.entries(quality_by_domain).forEach(([domainId, domainData]) => { console.log(`${domainData.domain_name}:`); console.log(` Average confidence: ${(domainData.average_confidence * 100).toFixed(1)}%`); console.log(` Total mappings: ${domainData.total_mappings.toLocaleString()}`); const dist = domainData.quality_distribution; console.log(` Quality distribution:`); console.log(` Excellent: ${dist.excellent.toLocaleString()}`); console.log(` Good: ${dist.good.toLocaleString()}`); console.log(` Moderate: ${dist.moderate.toLocaleString()}`); console.log(` Poor: ${dist.poor.toLocaleString()}`); if (domainData.most_reliable_vocabulary_pairs.length > 0) { console.log(` Most reliable pairs: ${domainData.most_reliable_vocabulary_pairs.slice(0, 3).join(', ')}`); } if (domainData.quality_challenges.length > 0) { console.log(` Quality challenges: ${domainData.quality_challenges.slice(0, 2).join(', ')}`); } }); // Display vocabulary pair quality console.log('\n=== Top Quality Vocabulary Pairs ==='); const vocabPairs = []; Object.entries(qualityData.data.quality_by_vocabulary_pair).forEach(([sourceVocab, sourceData]) => { Object.entries(sourceData.target_vocabularies).forEach(([targetVocab, pairData]) => { vocabPairs.push({ pair: `${sourceVocab} → ${targetVocab}`, confidence: pairData.average_confidence, mappings: pairData.total_mappings, rating: pairData.quality_rating }); }); }); vocabPairs.sort((a, b) => b.confidence - a.confidence) .slice(0, 10) .forEach(pair => { console.log(`${pair.pair}: ${(pair.confidence * 100).toFixed(1)}% avg confidence (${pair.mappings.toLocaleString()} mappings, ${pair.rating})`); }); // Display quality metrics if (quality_metrics) { console.log('\n=== Quality Metrics ==='); if (quality_metrics.confidence_score_distribution) { const confDist = quality_metrics.confidence_score_distribution; console.log('Confidence Score Statistics:'); console.log(` Mean: ${(qualityData.data.quality_overview.overall_average_confidence * 100).toFixed(1)}%`); if (confDist.percentiles) { console.log(` 25th percentile: ${(confDist.percentiles['25'] * 100).toFixed(1)}%`); console.log(` 50th percentile: ${(confDist.percentiles['50'] * 100).toFixed(1)}%`); console.log(` 75th percentile: ${(confDist.percentiles['75'] * 100).toFixed(1)}%`); } } if (quality_metrics.mapping_source_analysis) { console.log('\nQuality by Mapping Source:'); Object.entries(quality_metrics.mapping_source_analysis).forEach(([source, data]) => { if (data.average_confidence) { console.log(` ${source}: ${(data.average_confidence * 100).toFixed(1)}% avg confidence`); } }); } } // Display recommendations if (qualityData.data.recommendations) { console.log('\n=== Quality Improvement Recommendations ==='); qualityData.data.recommendations.slice(0, 5).forEach((rec, i) => { console.log(`${i + 1}. ${rec.title} (${rec.priority} priority)`); console.log(` ${rec.description}`); console.log(` Expected improvement: ${rec.expected_improvement}`); if (rec.affected_vocabulary_pairs.length > 0) { console.log(` Affects: ${rec.affected_vocabulary_pairs.slice(0, 3).join(', ')}`); } }); } ``` ```python Python theme={null} import requests source_vocab = "SNOMED" target_vocab = "ICD10CM" url = f"https://api.omophub.com/v1/mappings/quality/{source_vocab}/{target_vocab}" headers = { "Authorization": "Bearer YOUR_API_KEY" } response = requests.get(url, headers=headers) data = response.json() print("=== MAPPING QUALITY ANALYSIS ===") overview = data['data']['quality_overview'] print(f"Total mappings analyzed: {overview['total_mappings_analyzed']:,}") print(f"Overall quality rating: {overview['overall_quality_rating']}") print(f"Average confidence: {overview['overall_average_confidence']:.3f}") print(f"High-quality mappings: {overview['high_quality_percentage']:.1f}%") print(f"Validated mappings: {overview['validated_percentage']:.1f}%") print(f"Disputed mappings: {overview['disputed_percentage']:.1f}%") # Domain quality analysis print(f"\n=== QUALITY BY DOMAIN ===") domain_quality = data['data']['quality_by_domain'] domain_df = pd.DataFrame([ { 'domain': info['domain_name'], 'total_mappings': info['total_mappings'], 'avg_confidence': info['average_confidence'], 'excellent': info['quality_distribution']['excellent'], 'good': info['quality_distribution']['good'], 'moderate': info['quality_distribution']['moderate'], 'poor': info['quality_distribution']['poor'] } for domain_id, info in domain_quality.items() ]) domain_df = domain_df.sort_values('avg_confidence', ascending=False) print(domain_df.to_string(index=False, float_format='%.3f')) # Vocabulary pair quality analysis print(f"\n=== VOCABULARY PAIR QUALITY ===") vocab_pairs = [] for source_vocab, source_data in data['data']['quality_by_vocabulary_pair'].items(): for target_vocab, pair_data in source_data['target_vocabularies'].items(): vocab_pairs.append({ 'source': source_vocab, 'target': target_vocab, 'pair': f"{source_vocab} → {target_vocab}", 'avg_confidence': pair_data['average_confidence'], 'total_mappings': pair_data['total_mappings'], 'validation_rate': pair_data['validation_rate'], 'quality_rating': pair_data['quality_rating'] }) vocab_pair_df = pd.DataFrame(vocab_pairs) vocab_pair_df = vocab_pair_df.sort_values('avg_confidence', ascending=False) print("Top 10 Quality Vocabulary Pairs:") top_pairs = vocab_pair_df.head(10) for _, row in top_pairs.iterrows(): print(f"{row['pair']}: {row['avg_confidence']:.3f} confidence, {row['total_mappings']:,} mappings ({row['quality_rating']})") print("\nBottom 5 Quality Vocabulary Pairs:") bottom_pairs = vocab_pair_df.tail(5) for _, row in bottom_pairs.iterrows(): print(f"{row['pair']}: {row['avg_confidence']:.3f} confidence, {row['total_mappings']:,} mappings ({row['quality_rating']})") # Quality metrics analysis if 'quality_metrics' in data['data']: metrics = data['data']['quality_metrics'] print(f"\n=== DETAILED QUALITY METRICS ===") # Confidence distribution if 'confidence_score_distribution' in metrics: conf_dist = metrics['confidence_score_distribution'] print(f"Confidence Score Statistics:") if 'percentiles' in conf_dist: for percentile, value in conf_dist['percentiles'].items(): print(f" {percentile}th percentile: {value:.3f}") if 'skewness' in conf_dist: print(f" Skewness: {conf_dist['skewness']:.3f}") if 'kurtosis' in conf_dist: print(f" Kurtosis: {conf_dist['kurtosis']:.3f}") # Semantic similarity analysis if 'semantic_similarity_analysis' in metrics: sem_analysis = metrics['semantic_similarity_analysis'] print(f"\nSemantic Similarity Analysis:") print(f" Average similarity: {sem_analysis['average_similarity']:.3f}") print(f" Similarity-confidence correlation: {sem_analysis['similarity_confidence_correlation']:.3f}") print(f" Low similarity, high confidence: {sem_analysis['low_similarity_high_confidence']:,}") print(f" High similarity, low confidence: {sem_analysis['high_similarity_low_confidence']:,}") # Mapping source analysis if 'mapping_source_analysis' in metrics: source_analysis = metrics['mapping_source_analysis'] print(f"\nQuality by Mapping Source:") for source, source_data in source_analysis.items(): if 'average_confidence' in source_data: print(f" {source}: {source_data['average_confidence']:.3f} avg confidence") # Outlier analysis if 'quality_outliers' in data['data']: outliers = data['data']['quality_outliers'] print(f"\n=== QUALITY OUTLIERS ===") if outliers.get('low_quality_outliers'): print(f"Low Quality Outliers ({len(outliers['low_quality_outliers'])}):") for outlier in outliers['low_quality_outliers'][:5]: print(f" Concepts {outlier['source_concept_id']} → {outlier['target_concept_id']}") print(f" Confidence: {outlier['confidence_score']:.3f}") print(f" Reason: {outlier['outlier_reason']}") if outliers.get('outlier_patterns'): print(f"\nCommon Outlier Patterns:") for pattern in outliers['outlier_patterns'][:3]: print(f" - {pattern}") # Trends analysis if 'quality_trends' in data['data']: trends = data['data']['quality_trends'] print(f"\n=== QUALITY TRENDS ===") print(f"Quality improvement rate: {trends['quality_improvement_rate']:.2f}% annually") if trends.get('fastest_improving_areas'): print(f"Fastest improving areas:") for area in trends['fastest_improving_areas'][:3]: print(f" - {area}") # Plot quality trend if trends.get('quality_over_time'): quality_history = trends['quality_over_time'] dates = [point['date'] for point in quality_history] confidences = [point['average_confidence'] for point in quality_history] validation_rates = [point['validation_rate'] for point in quality_history] fig, (ax1, ax2) = plt.subplots(2, 1, figsize=(12, 10)) # Confidence trend ax1.plot(dates, confidences, marker='o', linewidth=2) ax1.set_title('Average Confidence Score Over Time') ax1.set_ylabel('Confidence Score') ax1.tick_params(axis='x', rotation=45) ax1.grid(True, alpha=0.3) # Validation rate trend ax2.plot(dates, validation_rates, marker='s', linewidth=2, color='green') ax2.set_title('Validation Rate Over Time') ax2.set_xlabel('Date') ax2.set_ylabel('Validation Rate') ax2.tick_params(axis='x', rotation=45) ax2.grid(True, alpha=0.3) plt.tight_layout() plt.show() # Recommendations if 'recommendations' in data['data']: print(f"\n=== QUALITY IMPROVEMENT RECOMMENDATIONS ===") recommendations = data['data']['recommendations'] for i, rec in enumerate(recommendations[:7], 1): print(f"{i}. {rec['title']} ({rec['priority']} Priority)") print(f" {rec['description']}") print(f" Expected improvement: {rec['expected_improvement']}") print(f" Implementation effort: {rec['implementation_effort']}") if rec.get('affected_vocabulary_pairs'): print(f" Affects: {', '.join(rec['affected_vocabulary_pairs'][:3])}") print() # Create comprehensive quality dashboard fig, ((ax1, ax2), (ax3, ax4)) = plt.subplots(2, 2, figsize=(16, 12)) # 1. Domain quality comparison domain_names = domain_df['domain'] domain_confidences = domain_df['avg_confidence'] ax1.barh(domain_names, domain_confidences, color='skyblue') ax1.set_xlabel('Average Confidence Score') ax1.set_title('Quality by Medical Domain') ax1.set_xlim(0, 1) # 2. Quality distribution pie chart quality_counts = [ domain_df['excellent'].sum(), domain_df['good'].sum(), domain_df['moderate'].sum(), domain_df['poor'].sum() ] quality_labels = ['Excellent', 'Good', 'Moderate', 'Poor'] colors = ['#2ecc71', '#3498db', '#f39c12', '#e74c3c'] ax2.pie(quality_counts, labels=quality_labels, colors=colors, autopct='%1.1f%%') ax2.set_title('Overall Quality Distribution') # 3. Vocabulary pair quality heatmap if len(vocab_pair_df) > 0: # Create pivot table for heatmap pivot_data = vocab_pair_df.pivot_table( values='avg_confidence', index='source', columns='target', fill_value=0 ) sns.heatmap(pivot_data, annot=True, fmt='.2f', cmap='RdYlGn', ax=ax3) ax3.set_title('Vocabulary Pair Quality Matrix') ax3.set_xlabel('Target Vocabulary') ax3.set_ylabel('Source Vocabulary') # 4. Confidence score histogram if 'quality_metrics' in data['data'] and 'confidence_score_distribution' in data['data']['quality_metrics']: conf_dist = data['data']['quality_metrics']['confidence_score_distribution'] if 'histogram_bins' in conf_dist: bins = conf_dist['histogram_bins'] bin_centers = [(bin['min'] + bin['max']) / 2 for bin in bins] bin_counts = [bin['count'] for bin in bins] ax4.bar(bin_centers, bin_counts, width=0.05, alpha=0.7, color='lightcoral') ax4.set_xlabel('Confidence Score') ax4.set_ylabel('Number of Mappings') ax4.set_title('Confidence Score Distribution') ax4.set_xlim(0, 1) plt.tight_layout() plt.show() # Summary report print(f"\n=== SUMMARY REPORT ===") print(f"• Analyzed {overview['total_mappings_analyzed']:,} mappings across {len(domain_quality)} domains") print(f"• Overall system quality rating: {overview['overall_quality_rating'].upper()}") print(f"• {overview['high_quality_percentage']:.1f}% of mappings are high-quality (confidence > 0.8)") print(f"• {overview['validated_percentage']:.1f}% of mappings are officially validated") print(f"• Top quality domain: {domain_df.iloc[0]['domain']} ({domain_df.iloc[0]['avg_confidence']:.3f} avg confidence)") print(f"• Best vocabulary pair: {vocab_pair_df.iloc[0]['pair']} ({vocab_pair_df.iloc[0]['avg_confidence']:.3f} confidence)") if 'recommendations' in data['data']: high_priority_recs = [r for r in data['data']['recommendations'] if r['priority'] == 'high'] print(f"• {len(high_priority_recs)} high-priority improvement recommendations identified") ``` ```json theme={null} { "success": true, "data": { "quality_overview": { "total_mappings_analyzed": 4578923, "mappings_with_quality_data": 3456789, "overall_average_confidence": 0.847, "overall_quality_rating": "good", "high_quality_percentage": 72.3, "validated_percentage": 68.9, "disputed_percentage": 2.7 }, "quality_by_vocabulary_pair": { "SNOMED": { "vocabulary_name": "Systematized Nomenclature of Medicine Clinical Terms", "target_vocabularies": { "ICD10CM": { "target_vocabulary_name": "International Classification of Diseases, Tenth Revision, Clinical Modification", "total_mappings": 156789, "average_confidence": 0.923, "median_confidence": 0.950, "confidence_std_dev": 0.087, "equivalence_distribution": { "exact": 89234, "broader": 45678, "narrower": 12345, "related": 9532 }, "validation_rate": 0.847, "quality_rating": "excellent", "primary_use_cases": ["billing", "quality_reporting", "clinical_documentation"] }, "HCPCS": { "target_vocabulary_name": "Healthcare Common Procedure Coding System", "total_mappings": 67890, "average_confidence": 0.756, "median_confidence": 0.780, "confidence_std_dev": 0.145, "equivalence_distribution": { "exact": 23456, "broader": 34567, "narrower": 6789, "related": 3078 }, "validation_rate": 0.623, "quality_rating": "good", "primary_use_cases": ["billing", "procedure_coding"] } }, "overall_outgoing_quality": 0.875 } }, "quality_by_domain": { "Condition": { "domain_name": "Condition", "total_mappings": 2847562, "average_confidence": 0.891, "quality_distribution": { "excellent": 1456789, "good": 987654, "moderate": 345678, "poor": 57441 }, "most_reliable_vocabulary_pairs": [ "SNOMED → ICD10CM", "ICD10CM → SNOMED", "SNOMED → ICD10" ], "least_reliable_vocabulary_pairs": [ "Read → ICPC2", "Local → Standard" ], "quality_challenges": [ "Complex conditions often map to broader categories", "Regional terminology variations affect consistency", "Rare diseases have limited mapping coverage" ] }, "Drug": { "domain_name": "Drug", "total_mappings": 987654, "average_confidence": 0.823, "quality_distribution": { "excellent": 456789, "good": 345678, "moderate": 123456, "poor": 61731 }, "most_reliable_vocabulary_pairs": [ "RxNorm → NDC", "NDC → RxNorm" ], "least_reliable_vocabulary_pairs": [ "SNOMED → RxNorm", "Local drug codes → RxNorm" ], "quality_challenges": [ "Generic vs brand name mapping complexity", "Dosage form variations affect precision", "Discontinued medications create gaps" ] } }, "quality_metrics": { "confidence_score_distribution": { "histogram_bins": [ {"min": 0.0, "max": 0.1, "count": 12456}, {"min": 0.1, "max": 0.2, "count": 23456}, {"min": 0.8, "max": 0.9, "count": 456789}, {"min": 0.9, "max": 1.0, "count": 678901} ], "percentiles": { "25": 0.734, "50": 0.847, "75": 0.923, "95": 0.978 }, "skewness": -0.342, "kurtosis": 2.156 }, "semantic_similarity_analysis": { "average_similarity": 0.823, "similarity_confidence_correlation": 0.756, "low_similarity_high_confidence": 23456, "high_similarity_low_confidence": 12345 }, "mapping_source_analysis": { "official_mappings": { "average_confidence": 0.912, "total_count": 2345678, "validation_rate": 0.923 }, "community_mappings": { "average_confidence": 0.734, "total_count": 456789, "validation_rate": 0.456 }, "algorithmic_mappings": { "average_confidence": 0.678, "total_count": 567890, "validation_rate": 0.234 } } }, "quality_outliers": { "low_quality_outliers": [ { "source_concept_id": 12345, "target_concept_id": 67890, "confidence_score": 0.123, "outlier_reason": "Very low confidence despite official mapping", "potential_issues": [ "Semantic mismatch", "Outdated mapping relationship" ] } ], "outlier_patterns": [ "Algorithmic mappings between distant concept classes", "Legacy mappings not updated with vocabulary revisions", "Cross-domain mappings with semantic drift" ] }, "recommendations": [ { "recommendation_type": "validation_review", "title": "Review disputed SNOMED to HCPCS mappings", "description": "Systematic review of 2,847 disputed mappings between SNOMED procedures and HCPCS codes to improve validation rate", "priority": "high", "expected_improvement": "Increase HCPCS mapping quality by 15-20%", "affected_vocabulary_pairs": ["SNOMED → HCPCS"], "implementation_effort": "moderate" }, { "recommendation_type": "algorithmic_improvement", "title": "Enhance semantic similarity algorithms for drug mappings", "description": "Improve algorithmic mapping quality between drug vocabularies using enhanced semantic similarity models", "priority": "medium", "expected_improvement": "Increase drug domain confidence by 10%", "affected_vocabulary_pairs": ["RxNorm → NDC", "SNOMED → RxNorm"], "implementation_effort": "high" } ] }, "meta": { "request_id": "req_quality_analysis_234567", "timestamp": "2024-12-22T10:00:00Z", "vocab_release": "2025.2", "analysis_scope": { "vocabularies_included": 23, "domains_analyzed": 15, "mapping_relationships_analyzed": 4578923 }, "confidence_calculation_method": "weighted_semantic_similarity_with_validation" } } ``` ## Usage Examples ### SNOMED to ICD10CM Quality Analyze mapping quality between SNOMED and ICD10CM: ```javascript theme={null} const quality = await fetch('/v1/mappings/quality/SNOMED/ICD10CM'); ``` ### RxNorm to NDC Quality Analyze drug vocabulary mapping quality: ```javascript theme={null} const drugQuality = await fetch('/v1/mappings/quality/RxNorm/NDC'); ``` ### With Specific Vocabulary Release Query quality metrics for a specific vocabulary version: ```javascript theme={null} const versionedQuality = await fetch('/v1/mappings/quality/SNOMED/HCPCS?vocab_release=2025.1'); ``` ## Related Endpoints * [Get Concept Mappings](/api-reference/mappings/get-concept-mappings) - Individual concept mapping quality * [Get Vocabulary Mappings](/api-reference/mappings/get-vocabulary-mappings) - Quality within vocabulary pairs * [Get Mapping Coverage](/api-reference/mappings/get-mapping-coverage) - Coverage vs quality analysis ## Notes * Quality analysis requires substantial computational resources for large datasets * Confidence scores are calculated using multiple factors including semantic similarity and validation status * Official mappings generally have higher quality than algorithmic or community mappings * Quality can vary significantly between vocabulary pairs and domains * Disputed mappings may indicate areas needing expert review * Quality trends help identify improvement or degradation over time * Outlier analysis reveals mappings that may need special attention * Quality thresholds help focus on the most reliable mappings for production use