Tutorials

Start with the Quickstart for one complete JAX/NumPyro analysis. These chapters explain each part of that workflow, in order. Use the API reference to look up a specific name or signature.

Step	What you will learn
1. Units and conventions	Prepare radii, velocities and errors; distinguish tracer number from mass
2. Choose a backend	Switch NumPy/JAX interfaces, measure execution time and use automatic differentiation
3. Define a model	Compose components, set parameters and implement a custom tracer
4. Predict observables	Evaluate densities, mass and LOS dispersion; plot and generate mock data
5. Run MCMC	Define priors, sample, check R-hat/ESS and NUTS divergences, and read corner contours
6. Save and resume	Export draws, preserve chain structure and restart the same analysis

Worked analyses

Apply the steps to complete spherical and axisymmetric analyses. The spherical example includes recorded outputs and their execution metadata. Both are short workflow examples; their completion does not demonstrate convergence or calibration.

• Spherical analysis with saved posteriors
• Axisymmetric synthetic analysis