Rubric Coverage
A point-by-point map of every grading rubric item to the page that delivers the evidence.
Univariate distributions for all numeric & categorical features. Bivariate conversion-rate charts by Income, Education, CCAvg, Family, CD Account. Pearson correlation chart with Personal_Loan.
Open EDADuplicate & missing-value checks, anomalous-Experience treatment, ID/ZIPCode drop with rationale, feature engineering (mortgage_flag, income_band, ccavg_band), stratified 70/30 split.
Open PreprocessingBaseline CART Decision Tree with classification metrics (Accuracy / Precision / Recall / F1 / AUC), confusion matrix, and explicit metric rationale for a campaign use case.
Open ModelingInteractive pre-pruning (max_depth, min_samples_split/leaf, criterion, class_weight) and post-pruning via cost-complexity α. Comparison table of baseline / pre-pruned / post-pruned with overfit gap. Final model selected with documented rationale and decision-rule export.
Open Model ImprovementSeven structured recommendations (Observation / Insight / Recommendation / Benefit / Limitation), tiered campaign strategy, operating cadence, and governance note.
Open RecommendationsExecutive-ready Professor Review page with rubric-aligned summary, hero numbers, feature importance, correlation chart, and downloadable PDF.
Open Professor Review