EV Purchase Predictor & Kaggle Classifier Studio

Interactive machine learning workbench replicating the Kaggle Playground (Season 6 Episode 9) CatBoost pipeline

Kaggle S6E9 Benchmark: ~94% ROC-AUC
Model Hyperparameters
CatBoost Simulator
Tree Depth (depth) 5
Controls tree complexity & interaction limits
Iterations (trees) 100
Boosting steps for gradient loss minimization
Learning Rate (η) 0.08
Shrinkage parameter scaling tree updates
Consumer Persona Evaluator
Live Scoring Engine
Representative Presets:
Annual Income ($k) $95k
Daily Commute (miles) 42 mi
Monthly Gas Expense ($) $280
Home Charging Access Yes
Green Values Index (1-10) 8 / 10
Govt Tax Credit Eligible Yes
Predicted Purchase Likelihood
91.2%
Will Buy EV (Probability: 91.2%)
Model Diagnostics
ROC-AUC: 0.941
ROC-AUC
0.941
Log Loss
0.246
Accuracy
89.4%
Receiver Operating Characteristic (ROC)
Feature Importance
Gini Importance

Top Driver: Home Charging Access provides the dominant split gain, followed by fuel displacement economics.

EDA Visualizer
Pie Chart (≤5 values)
Batch Test Set & Kaggle Submission Pipeline
Test Records Scored with Current Ensemble: 100 instances
ID Income ($k) Commute (mi) Gas ($/mo) Home Charger Green Score Rebate Pred Probability Status