`output_table_format`

Back to L7 | Browse all axes | Browse all options

Axis output_table_format on sub-layer L7_B_output_shape_export (layer l7).

Sub-layer

L7_B_output_shape_export

Axis metadata

Default: 'long'
Sweepable: False
Status: operational

Operational status summary

Operational: 2 option(s)
Future: 0 option(s)

Options

`long` – operational

Long-form (tidy) tables: one row per (model, feature, metric).

Returns importance tables in the tidy data format – each row is a single observation of (model_id, feature, metric_value). Default for downstream pandas / R analysis since aggregation, filtering, and ggplot-style faceting all expect this shape.

Wickham’s tidy-data principles (one variable per column, one observation per row, one type per table) underpin the long format.

When to use

Default for downstream pandas / R analysis; required for seaborn faceting.

When NOT to use

Paper-quality matrix-shaped reporting (use wide instead).

References

macroforecast design Part 3, L7: ‘every importance op produces (table, figure) pairs; the L7.B sub-layer governs export shape.’
Wickham (2014) ‘Tidy Data’, Journal of Statistical Software 59(10): 1-23. (doi:10.18637/jss.v059.i10)

Related options: wide

Last reviewed 2026-05-05 by macroforecast author.

`wide` – operational

Wide-form tables: one row per feature, columns per (model, metric).

Returns importance tables in the matrix-shaped format – each row is one feature, columns vary across (model_id × metric) combinations. Compact for paper-quality reporting and the natural shape for LaTeX tabular export.

When to use

Compact paper-quality reporting; LaTeX table generation.

When NOT to use

Downstream pandas analysis – use long instead.

References

macroforecast design Part 3, L7: ‘every importance op produces (table, figure) pairs; the L7.B sub-layer governs export shape.’

Related options: long