k-diagram: Polar Insights for Forecasting¶
Welcome to k-diagram! This package provides a unique perspective on evaluating forecasting models by leveraging the power of polar coordinates. Move beyond standard metrics and discover how circular plots can reveal deeper insights into your model’s performance, stability, and hidden weaknesses.
Ideal for: Data scientists, machine learning engineers, meteorologists, climate scientists, and researchers who need to diagnose and communicate the performance of complex forecasting models, especially when uncertainty is a key factor.
Start here to learn the core concepts and mathematical foundations behind the plots.
Browse every plot type with runnable code and interpretation guides.
Generate plots from your terminal. See the full list of commands and options.
Check out the latest features and fixes.
Analyze prediction intervals, coverage, anomalies, and drift.
Diagnose bias vs. variance, compare error distributions, visualize 2D uncertainty.
Generate Taylor diagrams and other evaluation views.
Visualize feature influence patterns (fingerprints).
Plot true vs. predicted values in polar coordinates.
Browse every function, class, and parameter.