Analyzer¶

Statistical analysis utilities for ASCI results.

Class Reference¶

`Analyzer` ¶

Analyzer for ASCI results

Functions¶

`init(results_df, config)` ¶

Initialize analyzer

PARAMETER	DESCRIPTION
`results_df`	ASCI calculation results TYPE: `DataFrame`
`config`	Reaction configuration TYPE: `ReactionConfig`

`get_top_catalysts(n=10, metric='ASCI')` ¶

Get top N catalysts by specified metric

PARAMETER	DESCRIPTION
`n`	Number of catalysts TYPE: `int` DEFAULT: `10`
`metric`	Metric to rank by ('ASCI', 'activity_score', etc.) TYPE: `str` DEFAULT: `'ASCI'`

RETURNS	DESCRIPTION
`DataFrame`	Top N catalysts

`get_pareto_optimal()` ¶

Get Pareto optimal catalysts

RETURNS	DESCRIPTION
`DataFrame`	Pareto optimal catalysts

`filter_by_threshold(metric, threshold, greater_than=True)` ¶

Filter catalysts by threshold

PARAMETER	DESCRIPTION
`metric`	Metric to filter by TYPE: `str`
`threshold`	Threshold value TYPE: `float`
`greater_than`	If True, keep values > threshold TYPE: `bool` DEFAULT: `True`

RETURNS	DESCRIPTION
`DataFrame`	Filtered catalysts

`get_statistics_by_element()` ¶

Get statistics grouped by element

RETURNS	DESCRIPTION
`dict`	Statistics by element

`get_correlation_analysis()` ¶

Get correlation matrix of key metrics

RETURNS	DESCRIPTION
`DataFrame`	Correlation matrix

`compare_weight_scenarios(weight_scenarios)` ¶

Compare different weight scenarios

PARAMETER	DESCRIPTION
`weight_scenarios`	List of (w_a, w_s, w_c) weight combinations TYPE: `list of tuples`

RETURNS	DESCRIPTION
`DataFrame`	Comparison results

`export_summary(output_path)` ¶

Export comprehensive summary

PARAMETER	DESCRIPTION
`output_path`	Output file path TYPE: `str`

Quick Reference¶

from ascicat.analyzer import Analyzer

analyzer = Analyzer(results, config)

# Basic statistics
stats = analyzer.get_statistics()

# Correlation analysis
corr = analyzer.compute_correlations()

# Distribution analysis
dist = analyzer.analyze_distribution()

Initialization¶

analyzer = Analyzer(
    results,    # DataFrame from calculate_asci()
    config      # ReactionConfig (optional)
)

Methods¶

get_statistics¶

Get summary statistics for all scores.

stats = analyzer.get_statistics()
print(stats['asci']['mean'])
print(stats['activity']['std'])

Returns: Nested dictionary with mean, std, min, max, median for each score.

compute_correlations¶

Compute correlations between scores.

corr_matrix = analyzer.compute_correlations()
print(corr_matrix)

Returns: DataFrame with Pearson correlations.

analyze_distribution¶

Analyze ASCI score distribution.

dist = analyzer.analyze_distribution()
print(f"Skewness: {dist['skewness']}")
print(f"Kurtosis: {dist['kurtosis']}")

Returns: Dictionary with distribution metrics.

identify_outliers¶

Find outlier catalysts.

outliers = analyzer.identify_outliers(method='iqr', threshold=1.5)
print(f"Found {len(outliers)} outliers")

compare_groups¶

Compare catalyst groups.

comparison = analyzer.compare_groups(
    group_col='Ametal',
    metric='ASCI'
)

Example Usage¶

from ascicat import ASCICalculator
from ascicat.analyzer import Analyzer

# Calculate ASCI
calc = ASCICalculator(reaction='HER')
calc.load_data('data/HER_clean.csv')
results = calc.calculate_asci()

# Analyze
analyzer = Analyzer(results, calc.config)

# Statistics
stats = analyzer.get_statistics()
print(f"ASCI: {stats['asci']['mean']:.3f} ± {stats['asci']['std']:.3f}")

# Correlations
corr = analyzer.compute_correlations()
print("\nScore Correlations:")
print(corr)

# Distribution
dist = analyzer.analyze_distribution()
print(f"\nSkewness: {dist['skewness']:.3f}")