coverage_view

Back to L1.5 | Browse all axes | Browse all options

Axis coverage_view on sub-layer L1_5_A_sample_coverage (layer l1_5).

Sub-layer

L1_5_A_sample_coverage

Axis metadata

  • Default: 'multi'

  • Sweepable: False

  • Status: operational

Operational status summary

  • Operational: 4 option(s)

  • Future: 0 option(s)

Options

multi – operational

Render every coverage view in a single composite output.

Composite view containing observation_count + per_series_start_end + panel_balance_matrix in one HTML / PDF report. Recommended default for exploratory data review.

When to use

First-pass exploratory data review covering all three coverage angles.

References

  • macroforecast design Part 4: ‘diagnostic layers default-off; non-blocking; produce JSON + matplotlib views attached to manifest.diagnostics/.’

Related options: observation_count, per_series_start_end, panel_balance_matrix

Last reviewed 2026-05-05 by macroforecast author.

observation_count – operational

Per-series observation count vs sample length.

Bar chart of n_obs per series over the active sample window. Highlights series that may be too short for the L4 estimator – a Lasso fit needs roughly n_obs > 2 × n_predictors, and short series violate that constraint silently.

When to use

First-pass sanity check that no predictor is mostly missing.

References

  • macroforecast design Part 4: ‘diagnostic layers default-off; non-blocking; produce JSON + matplotlib views attached to manifest.diagnostics/.’

Related options: per_series_start_end, panel_balance_matrix, multi

Last reviewed 2026-05-05 by macroforecast author.

panel_balance_matrix – operational

Binary observed/missing matrix over the full sample.

Heatmap with rows = series, columns = dates, cells = 1 (observed) or 0 (missing). Reveals structural breaks in coverage – e.g. a block of series that all start in 1990, or a block that disappears after 2008. The ragged-edge pattern is best understood here before L1.E’s sample-window rule trims the panel.

When to use

Visualising ragged-edge problems before applying L1.E sample-window rules.

References

  • macroforecast design Part 4: ‘diagnostic layers default-off; non-blocking; produce JSON + matplotlib views attached to manifest.diagnostics/.’

Related options: per_series_start_end

Last reviewed 2026-05-05 by macroforecast author.

per_series_start_end – operational

First / last observation date per series.

Table of per-series (first_valid_date, last_valid_date). Catches stale series that stopped publishing (last date too old) or new series with too short a history (first date too recent). Critical before applying the L1.E sample-window rule.

When to use

Diagnosing balanced-vs-unbalanced panel decisions in L1.E.

References

  • macroforecast design Part 4: ‘diagnostic layers default-off; non-blocking; produce JSON + matplotlib views attached to manifest.diagnostics/.’

Related options: observation_count, panel_balance_matrix

Last reviewed 2026-05-05 by macroforecast author.