forecast_scale_view
Back to L4.5 | Browse all axes | Browse all options
Axis
forecast_scale_viewon sub-layerL4_5_B_forecast_scale_view(layerl4_5).
Sub-layer
L4_5_B_forecast_scale_view
Axis metadata
Default:
'both_overlay'Sweepable: False
Status: operational
Operational status summary
Operational: 3 option(s)
Future: 0 option(s)
Options
back_transformed_only – operational
Plot forecasts back-transformed to the target’s level.
L4.5.B forecast scale view back_transformed_only.
This option configures the forecast_scale_view axis on the L4_5_B_forecast_scale_view sub-layer of L4.5; output is emitted under manifest.diagnostics/l4_5/L4_5_B_forecast_scale_view/ alongside the other selected views.
When to use
Default reporting view; matches the audience’s mental model of the target.
References
macroforecast design Part 4: ‘diagnostic layers default-off; non-blocking; produce JSON + matplotlib views attached to manifest.diagnostics/.’
Related options: transformed_only, both_overlay
Last reviewed 2026-05-05 by macroforecast author.
both_overlay – operational
Overlay both scales in side-by-side panels.
L4.5.B forecast scale view both_overlay.
This option configures the forecast_scale_view axis on the L4_5_B_forecast_scale_view sub-layer of L4.5; output is emitted under manifest.diagnostics/l4_5/L4_5_B_forecast_scale_view/ alongside the other selected views.
When to use
Comparing transformed vs level forecasts; useful when transformation effects matter.
References
macroforecast design Part 4: ‘diagnostic layers default-off; non-blocking; produce JSON + matplotlib views attached to manifest.diagnostics/.’
Related options: transformed_only, back_transformed_only
Last reviewed 2026-05-05 by macroforecast author.
transformed_only – operational
Plot forecasts in the transformed (model-internal) scale.
L4.5.B forecast scale view transformed_only.
This option configures the forecast_scale_view axis on the L4_5_B_forecast_scale_view sub-layer of L4.5; output is emitted under manifest.diagnostics/l4_5/L4_5_B_forecast_scale_view/ alongside the other selected views.
When to use
Inspecting model-native predictions; useful for diagnosing fit issues at the estimator’s scale.
References
macroforecast design Part 4: ‘diagnostic layers default-off; non-blocking; produce JSON + matplotlib views attached to manifest.diagnostics/.’
Related options: back_transformed_only, both_overlay
Last reviewed 2026-05-05 by macroforecast author.