Load Data
Quick View
One-click context: stat cards, correlation heatmap, column insights, and target analysis.
Preprocessing
Data Analysis
A self-contained HTML report you can open, print, or share.
Result Snapshots & Diff
vs
Train Model
Feature Health (Smarter Filtering)
Detects near-constant columns, high-cardinality ID/free-text columns, and highly correlated feature pairs (multicollinearity) so you can drop uninformative / redundant inputs before training.
Save / Load Model
Make Predictions
Batch Predictions
Explain Prediction & What-If
Tip: click the "Explain" button on any prediction row below to see its drivers.
Batch What-If
Re-Predict Production Table
Scores every row with the trained model, flags feature drift vs. the training reference (PSI), and highlights anomalous rows.
LLM Advisor
Database Schema (ER Diagram)
Column Types
Auto-Join & Load
Sample Load
Saved Queries
Schema Drift Profiling
vs
Relational Deep Features
Experiments & Champion
Anomaly Detection
Model Monitoring (Drift)
Reusable Pipeline Recipes
A recipe captures the current data source + preprocessing + training config so it can be reproduced later on the same or updated data.