Visualizer¶

High-quality figure generation.

Class Reference¶

`Visualizer` ¶

Visualization for ASCICat results.

Generates four key figure panels: - Panel A: 3D ASCI component space - Panel B: Rank vs adsorption energy (with quadratic trend) - Panel C: Volcano optimization landscape (ASCI contours) - Panel D: Top performers breakdown

For large datasets (>1000 catalysts), automatic stratified sampling is applied to ensure clear figures.

CRITICAL AXIS RULES: - All axis limits derived ONLY from sampled DataFrame - Rank axis is strictly 1 to N_sampled (no global ranks) - Energy axes reflect actual data range (not full 50k dataset) - Score axes normalized to [0,1] based on sampled data

PARAMETER	DESCRIPTION
`results`	ASCI calculation results from ASCICalculator TYPE: `DataFrame`
`config`	Reaction configuration object TYPE: `ReactionConfig`
`auto_sample`	Automatically sample large datasets (default: True) TYPE: `bool` DEFAULT: `True`
`n_samples`	Number of samples for visualization (default: 1000) TYPE: `int` DEFAULT: `1000`
`renormalize_scores`	Re-normalize scores after sampling for better 3D distribution (default: True) TYPE: `bool` DEFAULT: `True`

Examples:

>>> calc = ASCICalculator('HER')
>>> calc.load_data('data/HER_clean.csv')
>>> results = calc.calculate_asci()
>>> viz = Visualizer(results, calc.config)
>>> viz.generate_publication_figures('figures/')

Functions¶

`generate_publication_figures(output_dir='figures', dpi=600, formats=['png', 'pdf'])` ¶

Generate all four figure panels.

PARAMETER	DESCRIPTION
`output_dir`	Output directory for figures TYPE: `str` DEFAULT: `'figures'`
`dpi`	Resolution (default: 600) TYPE: `int` DEFAULT: `600`
`formats`	Output formats (default: ['png', 'pdf']) TYPE: `list` DEFAULT: `['png', 'pdf']`

`plot_3d_pareto_space(figsize=(5.5, 5), panel_label='a')` ¶

Panel A: 3D visualization of ASCI component space.

Shows the multi-objective optimization landscape with Activity, Stability, and Cost scores on three axes.

`plot_rank_vs_adsorption(figsize=(5.5, 5), panel_label='b')` ¶

Panel B: ASCI rank vs adsorption energy (volcano-style).

Shows the activity-performance relationship with the Sabatier volcano principle. Uses quadratic polynomial fit for the performance trend.

`plot_volcano_optimization(figsize=(5.5, 5), panel_label='c', n_contours=20)` ¶

Panel C: Volcano optimization landscape.

Shows ASCI contours over (ΔE, log₁₀ Cost) space, revealing the multi-objective optimization landscape.

CRITICAL: Computes ASCI using the formula on a grid, not interpolating raw values. Colorbar is 0-1 range.

`plot_top_performers(figsize=(6, 5), panel_label='d', n_top=8)` ¶

Panel D: Top performers breakdown.

Shows component score breakdown for top catalysts as grouped bar chart.

`plot_combined_figure(figsize=(12, 10), save_path=None, dpi=600)` ¶

Generate combined 4-panel figure.

`create_interactive_3d(output_path=None, auto_open=False)` ¶

Create interactive 3D visualization using Plotly.

`generate_all_outputs(output_dir='results', dpi=600, include_interactive=True)` ¶

Generate all outputs: static figures + interactive HTML.

Quick Reference¶

from ascicat.visualizer import Visualizer

viz = Visualizer(results, config, auto_sample=True)

# Generate all figures
viz.generate_all_outputs(output_dir='figures/')

# Individual plots
fig = viz.plot_3d_component_space()
fig = viz.plot_rank_vs_adsorption()
fig = viz.plot_volcano_optimization()
fig = viz.plot_top_performers()

Initialization¶

viz = Visualizer(
    results,           # DataFrame from calculate_asci()
    config,            # ReactionConfig from calculator
    auto_sample=True,  # Sample large datasets
    default_dpi=600,   # Output resolution
    color_palette='colorblind'  # Color scheme
)

Parameters:

Parameter	Type	Default	Description
`results`	DataFrame	Required	ASCI results
`config`	ReactionConfig	Required	Reaction configuration
`auto_sample`	bool	True	Sample datasets > 5000
`default_dpi`	int	600	Default resolution
`color_palette`	str	'colorblind'	Seaborn palette

Methods¶

generate_all_outputs¶

Generate complete figure set.

viz.generate_all_outputs(
    output_dir='figures/',
    dpi=600,
    formats=['png', 'pdf'],
    include_interactive=True
)

generate_figures¶

Generate four-panel figure.

viz.generate_figures(
    output_dir='figures/',
    dpi=600,
    formats=['png', 'pdf']
)

plot_3d_component_space¶

3D visualization of score space.

fig = viz.plot_3d_component_space(
    highlight_top_n=10,
    figsize=(8, 7),
    elev=25,
    azim=45
)

plot_rank_vs_adsorption¶

Activity volcano-style plot.

fig = viz.plot_rank_vs_adsorption(
    show_optimal=True,
    show_window=True,
    figsize=(8, 6)
)

plot_volcano_optimization¶

ASCI contour landscape.

fig = viz.plot_volcano_optimization(
    n_contours=20,
    cmap='plasma',
    figsize=(8, 6)
)

plot_top_performers¶

Score breakdown bar chart.

fig = viz.plot_top_performers(
    n_show=15,
    figsize=(10, 6)
)

plot_asci_distribution¶

ASCI score histogram.

fig = viz.plot_asci_distribution(
    bins=50,
    show_statistics=True
)

plot_score_correlations¶

Correlation matrix heatmap.

fig = viz.plot_score_correlations(figsize=(8, 7))

plot_pareto_2d¶

2D Pareto frontier projection.

fig = viz.plot_pareto_2d(
    x_score='activity',
    y_score='cost',
    highlight_top_n=20
)

plot_radar_chart¶

Spider/radar chart for catalyst comparison.

fig = viz.plot_radar_chart(
    catalyst_symbols=['Pt111', 'Fe2Sb4', 'Cu3Sb']
)

create_interactive_3d¶

Interactive Plotly 3D plot.

viz.create_interactive_3d(
    output_path='interactive.html',
    title='Catalyst Screening'
)

Output Formats¶

All static plots support:

PNG (raster, default 600 DPI)
PDF (vector)
SVG (vector)
EPS (vector)

fig.savefig('plot.png', dpi=600, bbox_inches='tight')
fig.savefig('plot.pdf', bbox_inches='tight')
fig.savefig('plot.svg', bbox_inches='tight')