--- name: forecast-accuracy-analyzer-950eda2ad4cd description: Forecast accuracy measurement and improvement skill with error decomposition allowed-tools: - Read - Write - Glob - Grep - Bash metadata: specialization: supply-chain domain: business category: analytics priority: standard graph: domains: [domain:supply-chain] specializations: [specialization:supply-chain-optimization] skillAreas: [skill-area:procurement-management, skill-area:statistical-analysis, skill-area:data-analytics] workflows: [workflow:vendor-onboarding, workflow:vendor-evaluation] roles: [role:supply-chain-analyst, role:data-analyst, role:procurement-manager] --- # Forecast Accuracy Analyzer ## Overview The Forecast Accuracy Analyzer provides comprehensive forecast accuracy measurement, error decomposition, and improvement recommendation capabilities. It supports continuous forecast quality improvement through root cause analysis and model performance comparison. ## Capabilities - **MAPE, WMAPE, Bias Calculation**: Standard accuracy metrics - **Forecast Error Decomposition**: Breakdown by error source - **SKU-Level Accuracy Tracking**: Granular accuracy monitoring - **Forecast Value-Add (FVA) Analysis**: Contribution of forecast steps - **Root Cause Categorization**: Error driver classification - **Model Performance Comparison**: Multi-model accuracy benchmarking - **Improvement Recommendation Generation**: Data-driven suggestions - **Accuracy Trend Monitoring**: Historical accuracy tracking ## Input Schema ```yaml forecast_accuracy_request: forecast_data: forecasts: array - sku_id: string period: string forecast_value: float forecast_source: string period_range: start: date end: date actual_data: actuals: array - sku_id: string period: string actual_value: float analysis_parameters: metrics: array # MAPE, WMAPE, Bias, etc. aggregation_levels: array # SKU, category, total fva_steps: array # Statistical, sales input, etc. segmentation: by_category: boolean by_volume: boolean by_variability: boolean ``` ## Output Schema ```yaml forecast_accuracy_output: accuracy_metrics: overall: mape: float wmape: float bias: float mpe: float by_segment: array by_sku: array error_decomposition: systematic_error: float random_error: float outlier_impact: float by_source: object fva_analysis: steps: array - step_name: string value_add: float before_accuracy: float after_accuracy: float recommendations: array root_cause_analysis: error_categories: array - category: string frequency: integer impact: float top_drivers: array model_comparison: models: array - model_name: string accuracy: float best_for: array improvement_recommendations: array - recommendation: string expected_improvement: float implementation_effort: string trends: accuracy_over_time: object bias_trend: object ``` ## Usage ### Monthly Accuracy Review ``` Input: Previous month's forecasts and actuals Process: Calculate accuracy metrics by segment Output: Accuracy report with performance analysis ``` ### Forecast Value-Add Analysis ``` Input: Forecast at each process step (statistical, sales, consensus) Process: Measure value added at each step Output: FVA report identifying low-value steps ``` ### Root Cause Investigation ``` Input: High-error SKUs, demand patterns Process: Categorize and analyze error drivers Output: Root cause report with recommendations ``` ## Integration Points - **Planning Systems**: Forecast and actual data - **BI Platforms**: Accuracy dashboards - **Statistical Tools**: Advanced analysis - **Tools/Libraries**: Statistical analysis, visualization ## Process Dependencies - Forecast Accuracy Analysis and Improvement - Demand Forecasting and Planning - Sales and Operations Planning (S&OP) ## Best Practices 1. Measure accuracy at multiple aggregation levels 2. Use weighted metrics for volume importance 3. Investigate outliers before concluding 4. Compare models on like-for-like basis 5. Set realistic improvement targets 6. Share accuracy results with stakeholders --- Source attribution: a5c-ai Original project: https://github.com/a5c-ai/babysitter License: MIT License information: https://opensource.org/license/mit Original upstream instructions; downloaded from Mega AI Hub. ## Upstream license notice MIT License Copyright (c) 2026 Gabriel Moreira Permission is hereby granted, free of charge, to any person obtaining a copy of this software and associated documentation files (the "Software"), to deal in the Software without restriction, including without limitation the rights to use, copy, modify, merge, publish, distribute, sublicense, and/or sell copies of the Software, and to permit persons to whom the Software is furnished to do so, subject to the following conditions: The above copyright notice and this permission notice shall be included in all copies or substantial portions of the Software. THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. # Mega AI Hub resource metadata Creator / publisher: a5c-ai Original project: https://github.com/a5c-ai/babysitter Attribution basis: Upstream repository explicitly recorded in package mirror path Fields of work: Data & Analytics, Marketing, Operations, Product Management Roles: Data Analyst, Growth Marketer, Market Research Analyst, Product Manager, Supply Chain Manager Specialties: API, Database Tools: Bi Platforms, Statistical Tools Country applicability: Country not specified Classification basis: Source catalog labels Instruction SHA-256: f3d5078601cf2b4db518083cc43f8c1136d53913a1453290d2c601ff8dc8bfa0