Data Science

knowledge skill

Performs exploratory data analysis and builds predictive models from tabular data. Produces a trained scikit-learn pipeline serialized as a .pkl file.

Worked examples

• {"input": {"data": "pd.DataFrame({'age':[25,30,35],'income':[50000,60000,70000],'purchased':[0,1,1]})", "target_column": "purchased", "model_type": "RandomForestClassifier"}, "output": {"model_artifact": "<bytes of sklearn.ensemble.RandomForestClassifier>", "metrics": {"accuracy": 0.95, "f1": 0.93}, "feature_importance": "pd.DataFrame({'feature':['income','age'], 'importance':[0.72,0.28]})"}}
• {"input": {"data": "pd.DataFrame({'sqft':[1200,1500,1800],'bedrooms':[2,3,3],'price':[250000,300000,320000]})", "target_column": "price", "model_type": "GradientBoostingRegressor"}, "output": {"model_artifact": "<bytes of sklearn.ensemble.GradientBoostingRegressor>", "metrics": {"rmse": 12500.0, "r2": 0.91}, "feature_importance": "pd.DataFrame({'feature':['sqft','bedrooms'], 'importance':[0.85,0.15]})"}}

Input

• data: DataFrame
• target_column: str
• model_type: str

Output

• model_artifact: bytes
• metrics: dict
• feature_importance: DataFrame

Details

• Skill type: knowledge skill
• Safety level: safe_public_research
• Version: 1.0.0