Start with your task
Use case first, reference second
Train a tabular model
Fit a structured regressor for taxi duration, fare, demand, or residual targets.
Choose a modeling path
Start with structured regression, panel forecasting, or advanced relationship models.
Model in the browser
Load a bundled sample and inspect structured regression or forecasting locally.
Validate the evidence
Read benchmark results by target, split, baseline, metric, and recommendation.
Modeling primitives
Built for place, time, and direction
- Periodic time splitters
- Spatial and route-aware trees
- Sparse zone memberships
- Leakage-aware validation
- Panel forecasting
- Portable model artifacts
First fit
A small model before the deep dive
# Run the complete dependency-free example:
python examples/quickstart.py
# Or start from the estimator directly:
from cartoboost import CartoBoostRegressor
model = CartoBoostRegressor(splitters=["axis", "periodic:24"])
model.fit(X_train, y_train)
predictions = model.predict(X_validation)Benchmark pages connect current-code results to split design, model settings, runtime, and what the result means for taxi-style modeling.
Read the benchmark reports