visualisation

visualisation

Module: visualisation.py

This module provides centralized visualization functions for the hybrid bias correction project. It generates diagnostic plots for two main domains:

QA Framework Visualization (from notebook 04 output): Eight plot types comparing correction quality across methods (LS, LSEQM, LSEQM+DL) and spatial domains for Indonesia.

Station Validation Visualization (from notebook 05 output): Spatial scatter maps, WMO multi-threshold performance curves, and regional box plots summarizing independent BMKG station validation results.

The eight plot types are: 1. CQI spatial maps – Continuous Quality Index for each method side by side. 2. Categorical quality maps – Poor / Fair / Good / Excellent classification. 3. Method improvement map – Quality gain from LS to LSEQM+DL. 4. Component quality maps – Basic statistical, distribution, and temporal scores. 5. Confidence map – Spatial reliability of the quality assessment. 6. CQI distribution analysis – Empirical CDF and histograms comparing methods. 7. Category summary – Grouped bar chart and percentage table. 8. Component box plots – Box plots comparing components across methods.

Each function can run interactively (plt.show) or in batch mode (plt.close). Figures are saved into per-plot-type sub-folders under the output directory, organized as {quality_prefix}_{plot_type}/ (e.g., qualitysd_cqi_spatial/).

The batch orchestrator, run_qa_batch_viz, loops over all 36 dekadal periods, calls all eight plot functions, and returns aggregated summary statistics. print_batch_summary displays the collected results.

Author: Benny Istanto Applied Climatology Study Program, Department of Geophysics and Meteorology, Bogor Agricultural University, Indonesia Email: bennyistanto@apps.ipb.ac.id

with supervision from Prof. Rizaldi Boer and Dr. I Putu Santikayasa

Update: 2026.03

Functions

Name	Description
load_quality_data	Load QA NetCDF files for all three correction methods.
plot_categorical_spatial	Categorical quality maps (Poor / Fair / Good / Excellent).
plot_category_summary	Grouped bar chart of quality-category percentages.
plot_component_boxplots	Box plots of component quality scores across methods.
plot_components	Component quality maps (basic, distribution, temporal) for best method.
plot_confidence	Confidence level map for the best available correction method.
plot_cqi_distribution	Empirical CDF and histogram of CQI across methods.
plot_cqi_spatial	CQI spatial maps - one panel per correction method.
plot_improvement	CQI improvement map (LSEQM+DL minus LS).
plot_multi_threshold_curves	WMO multi-threshold performance curves and exceedance frequency bar chart.
plot_qa_component_by_region	Box plots of QA component scores at stations, grouped by region.
plot_qa_province_bars	Horizontal bar chart of median CQI per province.
plot_qa_regional_bars	Grouped bar chart of median CQI per region for each correction method.
plot_qa_station_bars	Horizontal bar chart of CQI per individual station.
plot_regional_boxplots	Box plots of key validation metrics grouped by region or province.
plot_station_metric_maps	Spatial scatter maps of 6 key validation metrics at station locations.
print_batch_summary	Print aggregated statistics from a batch run.
print_station_validation_viz_summary	Print aggregated statistics from a station validation batch run.
run_qa_batch_viz	Generate all 8 plot types for all 36 dekadal periods.
run_qa_regional_batch_viz	Batch: QA regional/province/station plots for all 36 periods.
run_station_validation_batch_viz	Generate station validation plots for all 36 dekadal periods.

load_quality_data

visualisation.load_quality_data(
    month,
    dekad,
    quality_prefix='qualitysd',
    ref_label='cpc',
    config=None,
)

Load QA NetCDF files for all three correction methods.

Parameters

Name	Type	Description	Default
month	int	Month number (1–12).	required
dekad	int	Dekad number (1, 2, or 3).	required
quality_prefix	str	`'qualitysd'` (single-dekad aggregated) or `'qualityts'` (per-year timeseries).	`'qualitysd'`
ref_label	str	Reference dataset label (default `'cpc'`).	`'cpc'`
config	module	Configuration module. If None, imports `src.config`.	`None`

Returns

Name	Type	Description
	dict of {str: xarray.Dataset}	Keyed by display name (`'LS'`, `'LSEQM'`, `'LSEQMDL'`).

plot_categorical_spatial

visualisation.plot_categorical_spatial(
    quality_data,
    month,
    dekad,
    quality_prefix='qualitysd',
    output_dir=None,
    interactive=True,
)

Categorical quality maps (Poor / Fair / Good / Excellent).

plot_category_summary

visualisation.plot_category_summary(
    quality_data,
    month,
    dekad,
    quality_prefix='qualitysd',
    output_dir=None,
    interactive=True,
)

Grouped bar chart of quality-category percentages.

Returns

Name	Type	Description
fig	matplotlib.figure.Figure
summary	dict	`{method: {'counts': list, 'total': int, 'pcts': list}}`.

plot_component_boxplots

visualisation.plot_component_boxplots(
    quality_data,
    month,
    dekad,
    quality_prefix='qualitysd',
    output_dir=None,
    interactive=True,
)

Box plots of component quality scores across methods.

plot_components

visualisation.plot_components(
    quality_data,
    month,
    dekad,
    quality_prefix='qualitysd',
    output_dir=None,
    interactive=True,
)

Component quality maps (basic, distribution, temporal) for best method.

plot_confidence

visualisation.plot_confidence(
    quality_data,
    month,
    dekad,
    quality_prefix='qualitysd',
    output_dir=None,
    interactive=True,
)

Confidence level map for the best available correction method.

plot_cqi_distribution

visualisation.plot_cqi_distribution(
    quality_data,
    month,
    dekad,
    quality_prefix='qualitysd',
    output_dir=None,
    interactive=True,
)

Empirical CDF and histogram of CQI across methods.

plot_cqi_spatial

visualisation.plot_cqi_spatial(
    quality_data,
    month,
    dekad,
    quality_prefix='qualitysd',
    output_dir=None,
    interactive=True,
)

CQI spatial maps - one panel per correction method.

Parameters

Name	Type	Default
quality_data	dict of {str: xr.Dataset}	required
month	int	required
dekad	int	required
quality_prefix	str	`'qualitysd'`
output_dir	str or None	`None`
interactive	bool	`True`

Returns

Name	Type	Description
	matplotlib.figure.Figure

plot_improvement

visualisation.plot_improvement(
    quality_data,
    month,
    dekad,
    quality_prefix='qualitysd',
    output_dir=None,
    interactive=True,
)

CQI improvement map (LSEQM+DL minus LS).

Returns

Name	Type	Description
fig	matplotlib.figure.Figure
stats	dict or None	Improvement summary (mean, median, pct improved/degraded).

plot_multi_threshold_curves

visualisation.plot_multi_threshold_curves(
    all_mt_summaries,
    month,
    dekad,
    output_dir=None,
    interactive=True,
)

WMO multi-threshold performance curves and exceedance frequency bar chart.

Parameters

Name	Type	Description	Default
all_mt_summaries	dict of {str: pandas.DataFrame}	Keyed by method name. Each DataFrame is indexed by threshold (mm) with columns like `pod_median`, `pod_p25`, `pod_p75`, etc.	required
month	int	Period identifiers.	required
dekad	int	Period identifiers.	required
output_dir	str or None	If given, figures are saved into `{output_dir}/station_validation_threshold_curves/`.	`None`
interactive	bool	`True` → `plt.show()`; `False` → `plt.close()`.	`True`

Returns

Name	Type	Description
	matplotlib.figure.Figure or None

plot_qa_component_by_region

visualisation.plot_qa_component_by_region(
    quality_data,
    station_df,
    month,
    dekad,
    quality_prefix='qualitysd',
    method=None,
    output_dir=None,
    interactive=True,
)

Box plots of QA component scores at stations, grouped by region.

Four subplots (CQI, basic statistical, distribution, temporal) show the distribution of per-station QA values across the 7 island regions.

Parameters

Name	Type	Description	Default
quality_data	dict of {str: xr.Dataset}		required
station_df	pandas.DataFrame		required
month	int		required
dekad	int		required
quality_prefix	str		`'qualitysd'`
method	str or None	Method key (e.g. `'LS'`, `'LSEQM'`, `'LSEQMDL'`). If None, auto-picks best available.	`None`
output_dir	str or None		`None`
interactive	bool		`True`

Returns

Name	Type	Description
	matplotlib.figure.Figure or None

plot_qa_province_bars

visualisation.plot_qa_province_bars(
    quality_data,
    station_df,
    month,
    dekad,
    quality_prefix='qualitysd',
    method=None,
    output_dir=None,
    interactive=True,
)

Horizontal bar chart of median CQI per province.

Provinces are sorted by median CQI (descending). Station count and parent region are annotated beside each bar.

Parameters

Name	Type	Description	Default
quality_data	dict of {str: xr.Dataset}		required
station_df	pandas.DataFrame		required
month	int		required
dekad	int		required
quality_prefix	str		`'qualitysd'`
method	str or None	Method key (e.g. `'LS'`, `'LSEQM'`, `'LSEQMDL'`). If None, auto-picks best available.	`None`
output_dir	str or None		`None`
interactive	bool		`True`

Returns

Name	Type	Description
	matplotlib.figure.Figure or None

plot_qa_regional_bars

visualisation.plot_qa_regional_bars(
    quality_data,
    station_df,
    month,
    dekad,
    quality_prefix='qualitysd',
    output_dir=None,
    interactive=True,
)

Grouped bar chart of median CQI per region for each correction method.

X-axis shows the 7 island regions (from config.ISLAND_ORDER), with one bar per method (LS, LSEQM, LSEQM+DL). Station count is annotated above each group.

Parameters

Name	Type	Description	Default
quality_data	dict of {str: xr.Dataset}	Return of :func:`load_quality_data`.	required
station_df	pandas.DataFrame	Station metadata with `'ID_WMO'`, `'Lon'`, `'Lat'` columns.	required
month	int	Period identifiers.	required
dekad	int	Period identifiers.	required
quality_prefix	str	`'qualitysd'` or `'qualityts'`.	`'qualitysd'`
output_dir	str or None	If given, save into `{output_dir}/{quality_prefix}_qa_regional/`.	`None`
interactive	bool	`True` → `plt.show()`; `False` → `plt.close()`.	`True`

Returns

Name	Type	Description
	matplotlib.figure.Figure or None

plot_qa_station_bars

visualisation.plot_qa_station_bars(
    quality_data,
    station_df,
    month,
    dekad,
    quality_prefix='qualitysd',
    method=None,
    region_filter=None,
    output_dir=None,
    interactive=True,
)

Horizontal bar chart of CQI per individual station.

When region_filter is given, only stations in that region are plotted and a per-region figure is saved. Otherwise all stations are included in one (possibly large) figure.

Parameters

Name	Type	Description	Default
quality_data	dict of {str: xr.Dataset}		required
station_df	pandas.DataFrame		required
month	int		required
dekad	int		required
quality_prefix	str		`'qualitysd'`
method	str or None	Method key (e.g. `'LS'`, `'LSEQM'`, `'LSEQMDL'`). If None, auto-picks best available.	`None`
region_filter	str or None	If set, only plot stations from this region.	`None`
output_dir	str or None		`None`
interactive	bool		`True`

Returns

Name	Type	Description
	matplotlib.figure.Figure or None

plot_regional_boxplots

visualisation.plot_regional_boxplots(
    regional_df,
    month,
    dekad,
    method_name='LSEQMDL',
    group_col='Region',
    output_dir=None,
    interactive=True,
)

Box plots of key validation metrics grouped by region or province.

Parameters

Name	Type	Description	Default
regional_df	pandas.DataFrame	Per-station metrics with a `Region` or `Province` column (from `merge_station_metadata`).	required
month	int	Period identifiers.	required
dekad	int	Period identifiers.	required
method_name	str	Correction method label (for the plot title).	`'LSEQMDL'`
group_col	str	Column to group by (`'Region'` or `'Province'`).	`'Region'`
output_dir	str or None	If given, save to `{output_dir}/station_validation_regional/`.	`None`
interactive	bool	`True` → `plt.show()`; `False` → `plt.close()`.	`True`

Returns

Name	Type	Description
	matplotlib.figure.Figure or None

plot_station_metric_maps

visualisation.plot_station_metric_maps(
    metrics_df,
    station_df,
    month,
    dekad,
    method_name='LSEQMDL',
    output_dir=None,
    interactive=True,
)

Spatial scatter maps of 6 key validation metrics at station locations.

Parameters

Name	Type	Description	Default
metrics_df	pandas.DataFrame	Per-station metrics (index = WMO station ID).	required
station_df	pandas.DataFrame	Station locations with `ID_WMO`, `Lon`, `Lat` columns.	required
month	int	Period identifiers.	required
dekad	int	Period identifiers.	required
method_name	str	Correction method label (for the plot title).	`'LSEQMDL'`
output_dir	str or None	If given, save to `{output_dir}/station_validation_metric_maps/`.	`None`
interactive	bool	`True` → `plt.show()`; `False` → `plt.close()`.	`True`

Returns

Name	Type	Description
	matplotlib.figure.Figure

print_batch_summary

visualisation.print_batch_summary(summary)

Print aggregated statistics from a batch run.

Parameters

Name	Type	Description	Default
summary	dict	Return value of :func:`run_qa_batch_viz`.	required

print_station_validation_viz_summary

visualisation.print_station_validation_viz_summary(summary)

Print aggregated statistics from a station validation batch run.

Parameters

Name	Type	Description	Default
summary	dict	Return value of :func:`run_station_validation_batch_viz`.	required

run_qa_batch_viz

visualisation.run_qa_batch_viz(
    quality_prefix='qualitysd',
    output_dir=None,
    config=None,
    progress=True,
)

Generate all 8 plot types for all 36 dekadal periods.

Parameters

Name	Type	Description	Default
quality_prefix	str	`'qualitysd'` or `'qualityts'`.	`'qualitysd'`
output_dir	str or None	Base figures directory. Defaults to `{config.output_dir}/figures/qa`.	`None`
config	module or None		`None`
progress	bool	Print progress messages.	`True`

Returns

Name	Type	Description
	dict	`{(month, dekad): {'improvement': stats_dict, 'categories': summary_dict}}`

run_qa_regional_batch_viz

visualisation.run_qa_regional_batch_viz(
    quality_prefix='qualitysd',
    output_dir=None,
    config=None,
    progress=True,
)

Batch: QA regional/province/station plots for all 36 periods.

For each dekadal period, generates:

Regional grouped bar chart (CQI × method × 7 regions)
Component box plots by region (best method)
Province horizontal bars (best method)
Per-station bars for each region (best method)

Parameters

Name	Type	Description	Default
quality_prefix	str	`'qualitysd'` or `'qualityts'`.	`'qualitysd'`
output_dir	str or None	Base figures directory. Defaults to `{config.output_dir}/figures/qa`.	`None`
config	module or None		`None`
progress	bool	Print progress messages.	`True`

Returns

Name	Type	Description
	dict	`{(month, dekad): {'n_stations': int, 'n_methods': int}}`

run_station_validation_batch_viz

visualisation.run_station_validation_batch_viz(
    output_dir=None,
    config=None,
    progress=True,
)

Generate station validation plots for all 36 dekadal periods.

Reads CSV outputs from notebook 05 batch (station_validation/) and generates three plot types per period:

Station metric scatter maps (6 metrics × spatial)
WMO multi-threshold performance curves + exceedance bar chart
Regional box plots

Parameters

Name	Type	Description	Default
output_dir	str or None	Base figures directory. Defaults to `{config.output_dir}/figures/station_validation`.	`None`
config	module or None	Configuration module. If None, imports `src.config`.	`None`
progress	bool	Print progress messages.	`True`

Returns

Name	Type	Description
	dict	`{(month, dekad): {'n_methods': int, 'n_stations': int, ...}}`

# visualisation { #src.visualisation } `visualisation` Module: visualisation.py This module provides centralized visualization functions for the hybrid bias correction project. It generates diagnostic plots for two main domains: **QA Framework Visualization** (from notebook 04 output): Eight plot types comparing correction quality across methods (LS, LSEQM, LSEQM+DL) and spatial domains for Indonesia. **Station Validation Visualization** (from notebook 05 output): Spatial scatter maps, WMO multi-threshold performance curves, and regional box plots summarizing independent BMKG station validation results. The eight plot types are: 1. CQI spatial maps -- Continuous Quality Index for each method side by side. 2. Categorical quality maps -- Poor / Fair / Good / Excellent classification. 3. Method improvement map -- Quality gain from LS to LSEQM+DL. 4. Component quality maps -- Basic statistical, distribution, and temporal scores. 5. Confidence map -- Spatial reliability of the quality assessment. 6. CQI distribution analysis -- Empirical CDF and histograms comparing methods. 7. Category summary -- Grouped bar chart and percentage table. 8. Component box plots -- Box plots comparing components across methods. Each function can run interactively (plt.show) or in batch mode (plt.close). Figures are saved into per-plot-type sub-folders under the output directory, organized as {quality_prefix}_{plot_type}/ (e.g., qualitysd_cqi_spatial/). The batch orchestrator, run_qa_batch_viz, loops over all 36 dekadal periods, calls all eight plot functions, and returns aggregated summary statistics. print_batch_summary displays the collected results. **Author**: Benny Istanto Applied Climatology Study Program, Department of Geophysics and Meteorology, Bogor Agricultural University, Indonesia Email: bennyistanto@apps.ipb.ac.id with supervision from Prof. Rizaldi Boer and Dr. I Putu Santikayasa Update: 2026.03 ## Functions | Name | Description | | --- | --- | | [load_quality_data](#src.visualisation.load_quality_data) | Load QA NetCDF files for all three correction methods. | | [plot_categorical_spatial](#src.visualisation.plot_categorical_spatial) | Categorical quality maps (Poor / Fair / Good / Excellent). | | [plot_category_summary](#src.visualisation.plot_category_summary) | Grouped bar chart of quality-category percentages. | | [plot_component_boxplots](#src.visualisation.plot_component_boxplots) | Box plots of component quality scores across methods. | | [plot_components](#src.visualisation.plot_components) | Component quality maps (basic, distribution, temporal) for best method. | | [plot_confidence](#src.visualisation.plot_confidence) | Confidence level map for the best available correction method. | | [plot_cqi_distribution](#src.visualisation.plot_cqi_distribution) | Empirical CDF and histogram of CQI across methods. | | [plot_cqi_spatial](#src.visualisation.plot_cqi_spatial) | CQI spatial maps - one panel per correction method. | | [plot_improvement](#src.visualisation.plot_improvement) | CQI improvement map (LSEQM+DL minus LS). | | [plot_multi_threshold_curves](#src.visualisation.plot_multi_threshold_curves) | WMO multi-threshold performance curves and exceedance frequency bar chart. | | [plot_qa_component_by_region](#src.visualisation.plot_qa_component_by_region) | Box plots of QA component scores at stations, grouped by region. | | [plot_qa_province_bars](#src.visualisation.plot_qa_province_bars) | Horizontal bar chart of median CQI per province. | | [plot_qa_regional_bars](#src.visualisation.plot_qa_regional_bars) | Grouped bar chart of median CQI per region for each correction method. | | [plot_qa_station_bars](#src.visualisation.plot_qa_station_bars) | Horizontal bar chart of CQI per individual station. | | [plot_regional_boxplots](#src.visualisation.plot_regional_boxplots) | Box plots of key validation metrics grouped by region or province. | | [plot_station_metric_maps](#src.visualisation.plot_station_metric_maps) | Spatial scatter maps of 6 key validation metrics at station locations. | | [print_batch_summary](#src.visualisation.print_batch_summary) | Print aggregated statistics from a batch run. | | [print_station_validation_viz_summary](#src.visualisation.print_station_validation_viz_summary) | Print aggregated statistics from a station validation batch run. | | [run_qa_batch_viz](#src.visualisation.run_qa_batch_viz) | Generate all 8 plot types for all 36 dekadal periods. | | [run_qa_regional_batch_viz](#src.visualisation.run_qa_regional_batch_viz) | Batch: QA regional/province/station plots for all 36 periods. | | [run_station_validation_batch_viz](#src.visualisation.run_station_validation_batch_viz) | Generate station validation plots for all 36 dekadal periods. | ### load_quality_data { #src.visualisation.load_quality_data } ```python visualisation.load_quality_data( month, dekad, quality_prefix='qualitysd', ref_label='cpc', config=None, ) ``` Load QA NetCDF files for all three correction methods. #### Parameters {.doc-section .doc-section-parameters} | Name | Type | Description | Default | |----------------|--------------------|-------------------------------------------------------------------------------------|---------------| | month | [int](`int`) | Month number (1--12). | _required_ | | dekad | [int](`int`) | Dekad number (1, 2, or 3). | _required_ | | quality_prefix | [str](`str`) | ``'qualitysd'`` (single-dekad aggregated) or ``'qualityts'`` (per-year timeseries). | `'qualitysd'` | | ref_label | [str](`str`) | Reference dataset label (default ``'cpc'``). | `'cpc'` | | config | [module](`module`) | Configuration module. If *None*, imports ``src.config``. | `None` | #### Returns {.doc-section .doc-section-returns} | Name | Type | Description | |--------|-------------------------------|---------------------------------------------------------------| | | dict of {str: xarray.Dataset} | Keyed by display name (``'LS'``, ``'LSEQM'``, ``'LSEQMDL'``). | ### plot_categorical_spatial { #src.visualisation.plot_categorical_spatial } ```python visualisation.plot_categorical_spatial( quality_data, month, dekad, quality_prefix='qualitysd', output_dir=None, interactive=True, ) ``` Categorical quality maps (Poor / Fair / Good / Excellent). ### plot_category_summary { #src.visualisation.plot_category_summary } ```python visualisation.plot_category_summary( quality_data, month, dekad, quality_prefix='qualitysd', output_dir=None, interactive=True, ) ``` Grouped bar chart of quality-category percentages. #### Returns {.doc-section .doc-section-returns} | Name | Type | Description | |---------|-----------------------------------------------------------------------------------------------|-------------------------------------------------------------| | fig | [matplotlib](`matplotlib`).[figure](`matplotlib.figure`).[Figure](`matplotlib.figure.Figure`) | | | summary | [dict](`dict`) | ``{method: {'counts': list, 'total': int, 'pcts': list}}``. | ### plot_component_boxplots { #src.visualisation.plot_component_boxplots } ```python visualisation.plot_component_boxplots( quality_data, month, dekad, quality_prefix='qualitysd', output_dir=None, interactive=True, ) ``` Box plots of component quality scores across methods. ### plot_components { #src.visualisation.plot_components } ```python visualisation.plot_components( quality_data, month, dekad, quality_prefix='qualitysd', output_dir=None, interactive=True, ) ``` Component quality maps (basic, distribution, temporal) for best method. ### plot_confidence { #src.visualisation.plot_confidence } ```python visualisation.plot_confidence( quality_data, month, dekad, quality_prefix='qualitysd', output_dir=None, interactive=True, ) ``` Confidence level map for the best available correction method. ### plot_cqi_distribution { #src.visualisation.plot_cqi_distribution } ```python visualisation.plot_cqi_distribution( quality_data, month, dekad, quality_prefix='qualitysd', output_dir=None, interactive=True, ) ``` Empirical CDF and histogram of CQI across methods. ### plot_cqi_spatial { #src.visualisation.plot_cqi_spatial } ```python visualisation.plot_cqi_spatial( quality_data, month, dekad, quality_prefix='qualitysd', output_dir=None, interactive=True, ) ``` CQI spatial maps - one panel per correction method. #### Parameters {.doc-section .doc-section-parameters} | Name | Type | Description | Default | |----------------|---------------------------|---------------|---------------| | quality_data | dict of {str: xr.Dataset} | | _required_ | | month | [int](`int`) | | _required_ | | dekad | [int](`int`) | | _required_ | | quality_prefix | [str](`str`) | | `'qualitysd'` | | output_dir | [str](`str`) or None | | `None` | | interactive | [bool](`bool`) | | `True` | #### Returns {.doc-section .doc-section-returns} | Name | Type | Description | |--------|-----------------------------------------------------------------------------------------------|---------------| | | [matplotlib](`matplotlib`).[figure](`matplotlib.figure`).[Figure](`matplotlib.figure.Figure`) | | ### plot_improvement { #src.visualisation.plot_improvement } ```python visualisation.plot_improvement( quality_data, month, dekad, quality_prefix='qualitysd', output_dir=None, interactive=True, ) ``` CQI improvement map (LSEQM+DL minus LS). #### Returns {.doc-section .doc-section-returns} | Name | Type | Description | |--------|-----------------------------------------------------------------------------------------------|------------------------------------------------------------| | fig | [matplotlib](`matplotlib`).[figure](`matplotlib.figure`).[Figure](`matplotlib.figure.Figure`) | | | stats | [dict](`dict`) or None | Improvement summary (mean, median, pct improved/degraded). | ### plot_multi_threshold_curves { #src.visualisation.plot_multi_threshold_curves } ```python visualisation.plot_multi_threshold_curves( all_mt_summaries, month, dekad, output_dir=None, interactive=True, ) ``` WMO multi-threshold performance curves and exceedance frequency bar chart. #### Parameters {.doc-section .doc-section-parameters} | Name | Type | Description | Default | |------------------|---------------------------------|-------------------------------------------------------------------------------------------------------------------------------------|------------| | all_mt_summaries | dict of {str: pandas.DataFrame} | Keyed by method name. Each DataFrame is indexed by threshold (mm) with columns like ``pod_median``, ``pod_p25``, ``pod_p75``, etc. | _required_ | | month | [int](`int`) | Period identifiers. | _required_ | | dekad | [int](`int`) | Period identifiers. | _required_ | | output_dir | [str](`str`) or None | If given, figures are saved into ``{output_dir}/station_validation_threshold_curves/``. | `None` | | interactive | [bool](`bool`) | ``True`` → ``plt.show()``; ``False`` → ``plt.close()``. | `True` | #### Returns {.doc-section .doc-section-returns} | Name | Type | Description | |--------|-------------------------------------------------------------------------------------------------------|---------------| | | [matplotlib](`matplotlib`).[figure](`matplotlib.figure`).[Figure](`matplotlib.figure.Figure`) or None | | ### plot_qa_component_by_region { #src.visualisation.plot_qa_component_by_region } ```python visualisation.plot_qa_component_by_region( quality_data, station_df, month, dekad, quality_prefix='qualitysd', method=None, output_dir=None, interactive=True, ) ``` Box plots of QA component scores at stations, grouped by region. Four subplots (CQI, basic statistical, distribution, temporal) show the distribution of per-station QA values across the 7 island regions. #### Parameters {.doc-section .doc-section-parameters} | Name | Type | Description | Default | |----------------|----------------------------------------------------|-----------------------------------------------------------------------------------------------|---------------| | quality_data | dict of {str: xr.Dataset} | | _required_ | | station_df | [pandas](`pandas`).[DataFrame](`pandas.DataFrame`) | | _required_ | | month | [int](`int`) | | _required_ | | dekad | [int](`int`) | | _required_ | | quality_prefix | [str](`str`) | | `'qualitysd'` | | method | [str](`str`) or None | Method key (e.g. ``'LS'``, ``'LSEQM'``, ``'LSEQMDL'``). If *None*, auto-picks best available. | `None` | | output_dir | [str](`str`) or None | | `None` | | interactive | [bool](`bool`) | | `True` | #### Returns {.doc-section .doc-section-returns} | Name | Type | Description | |--------|-------------------------------------------------------------------------------------------------------|---------------| | | [matplotlib](`matplotlib`).[figure](`matplotlib.figure`).[Figure](`matplotlib.figure.Figure`) or None | | ### plot_qa_province_bars { #src.visualisation.plot_qa_province_bars } ```python visualisation.plot_qa_province_bars( quality_data, station_df, month, dekad, quality_prefix='qualitysd', method=None, output_dir=None, interactive=True, ) ``` Horizontal bar chart of median CQI per province. Provinces are sorted by median CQI (descending). Station count and parent region are annotated beside each bar. #### Parameters {.doc-section .doc-section-parameters} | Name | Type | Description | Default | |----------------|----------------------------------------------------|-----------------------------------------------------------------------------------------------|---------------| | quality_data | dict of {str: xr.Dataset} | | _required_ | | station_df | [pandas](`pandas`).[DataFrame](`pandas.DataFrame`) | | _required_ | | month | [int](`int`) | | _required_ | | dekad | [int](`int`) | | _required_ | | quality_prefix | [str](`str`) | | `'qualitysd'` | | method | [str](`str`) or None | Method key (e.g. ``'LS'``, ``'LSEQM'``, ``'LSEQMDL'``). If *None*, auto-picks best available. | `None` | | output_dir | [str](`str`) or None | | `None` | | interactive | [bool](`bool`) | | `True` | #### Returns {.doc-section .doc-section-returns} | Name | Type | Description | |--------|-------------------------------------------------------------------------------------------------------|---------------| | | [matplotlib](`matplotlib`).[figure](`matplotlib.figure`).[Figure](`matplotlib.figure.Figure`) or None | | ### plot_qa_regional_bars { #src.visualisation.plot_qa_regional_bars } ```python visualisation.plot_qa_regional_bars( quality_data, station_df, month, dekad, quality_prefix='qualitysd', output_dir=None, interactive=True, ) ``` Grouped bar chart of median CQI per region for each correction method. X-axis shows the 7 island regions (from ``config.ISLAND_ORDER``), with one bar per method (LS, LSEQM, LSEQM+DL). Station count is annotated above each group. #### Parameters {.doc-section .doc-section-parameters} | Name | Type | Description | Default | |----------------|----------------------------------------------------|---------------------------------------------------------------------|---------------| | quality_data | dict of {str: xr.Dataset} | Return of :func:`load_quality_data`. | _required_ | | station_df | [pandas](`pandas`).[DataFrame](`pandas.DataFrame`) | Station metadata with ``'ID_WMO'``, ``'Lon'``, ``'Lat'`` columns. | _required_ | | month | [int](`int`) | Period identifiers. | _required_ | | dekad | [int](`int`) | Period identifiers. | _required_ | | quality_prefix | [str](`str`) | ``'qualitysd'`` or ``'qualityts'``. | `'qualitysd'` | | output_dir | [str](`str`) or None | If given, save into ``{output_dir}/{quality_prefix}_qa_regional/``. | `None` | | interactive | [bool](`bool`) | ``True`` → ``plt.show()``; ``False`` → ``plt.close()``. | `True` | #### Returns {.doc-section .doc-section-returns} | Name | Type | Description | |--------|-------------------------------------------------------------------------------------------------------|---------------| | | [matplotlib](`matplotlib`).[figure](`matplotlib.figure`).[Figure](`matplotlib.figure.Figure`) or None | | ### plot_qa_station_bars { #src.visualisation.plot_qa_station_bars } ```python visualisation.plot_qa_station_bars( quality_data, station_df, month, dekad, quality_prefix='qualitysd', method=None, region_filter=None, output_dir=None, interactive=True, ) ``` Horizontal bar chart of CQI per individual station. When *region_filter* is given, only stations in that region are plotted and a per-region figure is saved. Otherwise all stations are included in one (possibly large) figure. #### Parameters {.doc-section .doc-section-parameters} | Name | Type | Description | Default | |----------------|----------------------------------------------------|-----------------------------------------------------------------------------------------------|---------------| | quality_data | dict of {str: xr.Dataset} | | _required_ | | station_df | [pandas](`pandas`).[DataFrame](`pandas.DataFrame`) | | _required_ | | month | [int](`int`) | | _required_ | | dekad | [int](`int`) | | _required_ | | quality_prefix | [str](`str`) | | `'qualitysd'` | | method | [str](`str`) or None | Method key (e.g. ``'LS'``, ``'LSEQM'``, ``'LSEQMDL'``). If *None*, auto-picks best available. | `None` | | region_filter | [str](`str`) or None | If set, only plot stations from this region. | `None` | | output_dir | [str](`str`) or None | | `None` | | interactive | [bool](`bool`) | | `True` | #### Returns {.doc-section .doc-section-returns} | Name | Type | Description | |--------|-------------------------------------------------------------------------------------------------------|---------------| | | [matplotlib](`matplotlib`).[figure](`matplotlib.figure`).[Figure](`matplotlib.figure.Figure`) or None | | ### plot_regional_boxplots { #src.visualisation.plot_regional_boxplots } ```python visualisation.plot_regional_boxplots( regional_df, month, dekad, method_name='LSEQMDL', group_col='Region', output_dir=None, interactive=True, ) ``` Box plots of key validation metrics grouped by region or province. #### Parameters {.doc-section .doc-section-parameters} | Name | Type | Description | Default | |-------------|----------------------------------------------------|-------------------------------------------------------------------------------------------------|-------------| | regional_df | [pandas](`pandas`).[DataFrame](`pandas.DataFrame`) | Per-station metrics with a ``Region`` or ``Province`` column (from ``merge_station_metadata``). | _required_ | | month | [int](`int`) | Period identifiers. | _required_ | | dekad | [int](`int`) | Period identifiers. | _required_ | | method_name | [str](`str`) | Correction method label (for the plot title). | `'LSEQMDL'` | | group_col | [str](`str`) | Column to group by (``'Region'`` or ``'Province'``). | `'Region'` | | output_dir | [str](`str`) or None | If given, save to ``{output_dir}/station_validation_regional/``. | `None` | | interactive | [bool](`bool`) | ``True`` → ``plt.show()``; ``False`` → ``plt.close()``. | `True` | #### Returns {.doc-section .doc-section-returns} | Name | Type | Description | |--------|-------------------------------------------------------------------------------------------------------|---------------| | | [matplotlib](`matplotlib`).[figure](`matplotlib.figure`).[Figure](`matplotlib.figure.Figure`) or None | | ### plot_station_metric_maps { #src.visualisation.plot_station_metric_maps } ```python visualisation.plot_station_metric_maps( metrics_df, station_df, month, dekad, method_name='LSEQMDL', output_dir=None, interactive=True, ) ``` Spatial scatter maps of 6 key validation metrics at station locations. #### Parameters {.doc-section .doc-section-parameters} | Name | Type | Description | Default | |-------------|----------------------------------------------------|---------------------------------------------------------------------|-------------| | metrics_df | [pandas](`pandas`).[DataFrame](`pandas.DataFrame`) | Per-station metrics (index = WMO station ID). | _required_ | | station_df | [pandas](`pandas`).[DataFrame](`pandas.DataFrame`) | Station locations with ``ID_WMO``, ``Lon``, ``Lat`` columns. | _required_ | | month | [int](`int`) | Period identifiers. | _required_ | | dekad | [int](`int`) | Period identifiers. | _required_ | | method_name | [str](`str`) | Correction method label (for the plot title). | `'LSEQMDL'` | | output_dir | [str](`str`) or None | If given, save to ``{output_dir}/station_validation_metric_maps/``. | `None` | | interactive | [bool](`bool`) | ``True`` → ``plt.show()``; ``False`` → ``plt.close()``. | `True` | #### Returns {.doc-section .doc-section-returns} | Name | Type | Description | |--------|-----------------------------------------------------------------------------------------------|---------------| | | [matplotlib](`matplotlib`).[figure](`matplotlib.figure`).[Figure](`matplotlib.figure.Figure`) | | ### print_batch_summary { #src.visualisation.print_batch_summary } ```python visualisation.print_batch_summary(summary) ``` Print aggregated statistics from a batch run. #### Parameters {.doc-section .doc-section-parameters} | Name | Type | Description | Default | |---------|----------------|-------------------------------------------|------------| | summary | [dict](`dict`) | Return value of :func:`run_qa_batch_viz`. | _required_ | ### print_station_validation_viz_summary { #src.visualisation.print_station_validation_viz_summary } ```python visualisation.print_station_validation_viz_summary(summary) ``` Print aggregated statistics from a station validation batch run. #### Parameters {.doc-section .doc-section-parameters} | Name | Type | Description | Default | |---------|----------------|-----------------------------------------------------------|------------| | summary | [dict](`dict`) | Return value of :func:`run_station_validation_batch_viz`. | _required_ | ### run_qa_batch_viz { #src.visualisation.run_qa_batch_viz } ```python visualisation.run_qa_batch_viz( quality_prefix='qualitysd', output_dir=None, config=None, progress=True, ) ``` Generate all 8 plot types for all 36 dekadal periods. #### Parameters {.doc-section .doc-section-parameters} | Name | Type | Description | Default | |----------------|----------------------------|--------------------------------------------------------------------------|---------------| | quality_prefix | [str](`str`) | ``'qualitysd'`` or ``'qualityts'``. | `'qualitysd'` | | output_dir | [str](`str`) or None | Base figures directory. Defaults to ``{config.output_dir}/figures/qa``. | `None` | | config | [module](`module`) or None | | `None` | | progress | [bool](`bool`) | Print progress messages. | `True` | #### Returns {.doc-section .doc-section-returns} | Name | Type | Description | |--------|----------------|-------------------------------------------------------------------------------| | | [dict](`dict`) | ``{(month, dekad): {'improvement': stats_dict, 'categories': summary_dict}}`` | ### run_qa_regional_batch_viz { #src.visualisation.run_qa_regional_batch_viz } ```python visualisation.run_qa_regional_batch_viz( quality_prefix='qualitysd', output_dir=None, config=None, progress=True, ) ``` Batch: QA regional/province/station plots for all 36 periods. For each dekadal period, generates: 1. Regional grouped bar chart (CQI × method × 7 regions) 2. Component box plots by region (best method) 3. Province horizontal bars (best method) 4. Per-station bars for each region (best method) #### Parameters {.doc-section .doc-section-parameters} | Name | Type | Description | Default | |----------------|----------------------------|--------------------------------------------------------------------------|---------------| | quality_prefix | [str](`str`) | ``'qualitysd'`` or ``'qualityts'``. | `'qualitysd'` | | output_dir | [str](`str`) or None | Base figures directory. Defaults to ``{config.output_dir}/figures/qa``. | `None` | | config | [module](`module`) or None | | `None` | | progress | [bool](`bool`) | Print progress messages. | `True` | #### Returns {.doc-section .doc-section-returns} | Name | Type | Description | |--------|----------------|-------------------------------------------------------------| | | [dict](`dict`) | ``{(month, dekad): {'n_stations': int, 'n_methods': int}}`` | ### run_station_validation_batch_viz { #src.visualisation.run_station_validation_batch_viz } ```python visualisation.run_station_validation_batch_viz( output_dir=None, config=None, progress=True, ) ``` Generate station validation plots for all 36 dekadal periods. Reads CSV outputs from notebook 05 batch (``station_validation/``) and generates three plot types per period: 1. Station metric scatter maps (6 metrics × spatial) 2. WMO multi-threshold performance curves + exceedance bar chart 3. Regional box plots #### Parameters {.doc-section .doc-section-parameters} | Name | Type | Description | Default | |------------|----------------------------|------------------------------------------------------------------------------------------|-----------| | output_dir | [str](`str`) or None | Base figures directory. Defaults to ``{config.output_dir}/figures/station_validation``. | `None` | | config | [module](`module`) or None | Configuration module. If *None*, imports ``src.config``. | `None` | | progress | [bool](`bool`) | Print progress messages. | `True` | #### Returns {.doc-section .doc-section-returns} | Name | Type | Description | |--------|----------------|------------------------------------------------------------------| | | [dict](`dict`) | ``{(month, dekad): {'n_methods': int, 'n_stations': int, ...}}`` |