Nixtla Benchmark Reporter

Purpose

Generate production-ready benchmark reports from forecasting accuracy metrics, enabling systematic model comparison and regression detection for Nixtla forecasting workflows.

Overview

This skill transforms raw forecast metrics (sMAPE, MASE, MAE, RMSE) into actionable insights. It:

• Parses benchmark results CSV files from statsforecast/TimeGPT experiments
• Calculates summary statistics (mean, median, std dev, percentiles)
• Generates model comparison tables with winners highlighted
• Creates regression detection reports comparing current vs. baseline results
• Produces GitHub issue templates for performance degradations
• Generates markdown reports with embedded charts and recommendations