Diagnostics Plots

diagnosticsHeatmapPlot — Multi-metric heatmap

Seven diagnostic metrics for all variables in a color-coded grid. Green = good, yellow = borderline, red = problematic.

plots.diagnosticsHeatmapPlot(element, data, opts);
plots.getDiagnosticsHeatmapData(data, opts);

Column	Good	Borderline	Bad
R-hat	< 1.01	< 1.05	≥ 1.05
Split R-hat	< 1.01	< 1.05	≥ 1.05
ESS/draw	> 0.1	> 0.05	≤ 0.05
Bulk ESS	> 400	> 100	≤ 100
Tail ESS	> 400	> 100	≤ 100
MCSE/SD	< 0.1	< 0.2	≥ 0.2
|Geweke z|	< 2	< 3	≥ 3

chainIntervalsPlot — Per-chain intervals

Diamond marker for each chain's posterior mean, with 90% HDI error bars. Background shading shows the overall 90% HDI. Dashed line shows the overall posterior mean.

plots.chainIntervalsPlot(element, data, 'mu', opts);

energyPlot — HMC energy diagnostics

Compares marginal energy distribution vs energy transition distribution as overlapping histograms. The E-BFMI (Bayesian Fraction of Missing Information) can be read visually: if the two histograms have very different widths, the HMC momentum is poorly tuned.

plots.energyPlot(element, data, opts);

Auto-detects energy variable in order: energy__, energy, lp__, log_density. Shows a message if no energy variable is found.

summaryTable — HTML table

Renders an HTML <table> (not a Plotly chart) directly into an element. 16 columns with color-coded cells.

plots.summaryTable(element, data, opts);

Columns: variable, mean, std, mcse, q5, q25, q50, q75, q95, ess, ess/draw, bulkEss, tailEss, rhat, splitRhat, |geweke z|, hdi90.

Cells are red if the value exceeds a diagnostic threshold, green if it passes. Supports dark/light theme via opts.theme.