factor_view
Back to L3.5 | Browse all axes | Browse all options
Axis
factor_viewon sub-layerL3_5_B_factor_block_inspection(layerl3_5).
Sub-layer
L3_5_B_factor_block_inspection
Axis metadata
Default:
'multi'Sweepable: False
Status: operational
Operational status summary
Operational: 5 option(s)
Future: 0 option(s)
Options
cumulative_variance – operational
Cumulative explained-variance curve.
L3.5.B factor view cumulative_variance.
This option configures the factor_view axis on the L3_5_B_factor_block_inspection sub-layer of L3.5; output is emitted under manifest.diagnostics/l3_5/L3_5_B_factor_block_inspection/ alongside the other selected views.
When to use
Quantifying how much variance the chosen n_components retains; threshold heuristics (80% / 90%) live here.
References
macroforecast design Part 4: ‘diagnostic layers default-off; non-blocking; produce JSON + matplotlib views attached to manifest.diagnostics/.’
Stock & Watson (2002) ‘Forecasting Using Principal Components from a Large Number of Predictors’, JASA 97(460): 1167-1179.
Related options: scree_plot, loadings_heatmap, factor_timeseries, multi
Last reviewed 2026-05-05 by macroforecast author.
factor_timeseries – operational
Estimated factor time-series plot.
L3.5.B factor view factor_timeseries.
This option configures the factor_view axis on the L3_5_B_factor_block_inspection sub-layer of L3.5; output is emitted under manifest.diagnostics/l3_5/L3_5_B_factor_block_inspection/ alongside the other selected views.
When to use
Confirming factors track recognisable cycles (NBER recessions, oil-price spikes, etc.).
References
macroforecast design Part 4: ‘diagnostic layers default-off; non-blocking; produce JSON + matplotlib views attached to manifest.diagnostics/.’
Stock & Watson (2002) ‘Forecasting Using Principal Components from a Large Number of Predictors’, JASA 97(460): 1167-1179.
Related options: scree_plot, loadings_heatmap, cumulative_variance, multi
Last reviewed 2026-05-05 by macroforecast author.
loadings_heatmap – operational
Heatmap of factor loadings (factors × predictors).
L3.5.B factor view loadings_heatmap.
This option configures the factor_view axis on the L3_5_B_factor_block_inspection sub-layer of L3.5; output is emitted under manifest.diagnostics/l3_5/L3_5_B_factor_block_inspection/ alongside the other selected views.
When to use
Interpreting factor identity; high-loading variables suggest the factor’s economic interpretation.
References
macroforecast design Part 4: ‘diagnostic layers default-off; non-blocking; produce JSON + matplotlib views attached to manifest.diagnostics/.’
Stock & Watson (2002) ‘Forecasting Using Principal Components from a Large Number of Predictors’, JASA 97(460): 1167-1179.
Related options: scree_plot, factor_timeseries, cumulative_variance, multi
Last reviewed 2026-05-05 by macroforecast author.
multi – operational
Render every factor view together.
L3.5.B factor view multi.
This option configures the factor_view axis on the L3_5_B_factor_block_inspection sub-layer of L3.5; output is emitted under manifest.diagnostics/l3_5/L3_5_B_factor_block_inspection/ alongside the other selected views.
When to use
Default rich factor diagnostic; the standard package for factor-model papers.
References
macroforecast design Part 4: ‘diagnostic layers default-off; non-blocking; produce JSON + matplotlib views attached to manifest.diagnostics/.’
Stock & Watson (2002) ‘Forecasting Using Principal Components from a Large Number of Predictors’, JASA 97(460): 1167-1179.
Related options: scree_plot, loadings_heatmap, factor_timeseries, cumulative_variance
Last reviewed 2026-05-05 by macroforecast author.
scree_plot – operational
Eigenvalue scree plot for PCA / SPCA / DFM blocks.
L3.5.B factor view scree_plot.
This option configures the factor_view axis on the L3_5_B_factor_block_inspection sub-layer of L3.5; output is emitted under manifest.diagnostics/l3_5/L3_5_B_factor_block_inspection/ alongside the other selected views.
When to use
Choosing n_components – the elbow heuristic remains the most popular tool.
References
macroforecast design Part 4: ‘diagnostic layers default-off; non-blocking; produce JSON + matplotlib views attached to manifest.diagnostics/.’
Stock & Watson (2002) ‘Forecasting Using Principal Components from a Large Number of Predictors’, JASA 97(460): 1167-1179.
Related options: loadings_heatmap, factor_timeseries, cumulative_variance, multi
Last reviewed 2026-05-05 by macroforecast author.