[bar-feature-importance] feature-importance: Feature Importance Bar Chart

[MarkusNeusinger]

Plot Request

Horizontal bar chart showing feature importances from ML models.

Use Cases

• Model interpretability
• Feature selection
• Understanding model decisions

Expected Features

• Horizontal bars sorted by importance
• Feature names on y-axis
• Importance values on x-axis
• Optional: error bars for ensemble methods
• Color gradient by importance

[claude]

✅ Spec ID: bar-feature-importance

Rationale: Horizontal bar chart showing ML feature importances is a distinct, specialized visualization combining horizontal bars with color gradients and optional error bars for model interpretability.

Add label approved to begin generation.

[github-actions]

✅ Specification Ready: bar-feature-importance

PR: #2281

---

specification.md

bar-feature-importance: Feature Importance Bar Chart

Description

A horizontal bar chart displaying feature importances from machine learning models, with features sorted by importance value and bars colored by a gradient to emphasize relative contribution. This visualization is essential for model interpretability, helping data scientists and stakeholders understand which features drive model predictions. The horizontal orientation allows for readable feature names of varying lengths.

Applications

• Explaining random forest or gradient boosting model decisions to stakeholders
• Feature selection by identifying low-importance variables for removal
• Model debugging by detecting unexpected high-importance features that may indicate data leakage

Data

• feature (categorical) - Names of the model features displayed on the y-axis
• importance (numeric) - Importance scores from the model, used for bar length
• std (numeric, optional) - Standard deviation of importance for ensemble methods, displayed as error bars
• Size: 10-30 features recommended for readability
• Example: Feature importances from sklearn RandomForestClassifier or XGBoost

Notes

• Sort bars by importance value (highest at top) for easy identification of key features
• Use a sequential color gradient (e.g., light to dark) mapped to importance values
• Error bars are optional but valuable for ensemble methods showing importance variability
• Consider showing only top N features if the model has many features
• Include importance values as text annotations at the end of bars for precision

---

Next: Add approved label to merge this specification to main.

---

🤖 spec-create workflow

[MarkusNeusinger]

✅ Closed with 9/9 implementations in repo (bar-feature-importance)