Guides¶
The guides say how to do one task each. The tutorials show whole analyses, and How it works has the maths.
An analysis has five steps, and each has a guide:
Prepare your inputs: turn your data into per-object likelihoods over bins (Prepare your inputs), and choose the bins (Choose bins).
Choose a prior for the population (Choose a prior). If your catalogue is incomplete, describe its selection (Selection and rates).
Run an engine (Choose an engine). The default chooses for you.
Read the result: population summaries, derived quantities and per-object posteriors (Read a result).
Check the result: diagnostics, then simulations (Check a result).
Scale up covers large problems. Troubleshooting maps each warning and symptom to its fix, and its pitfalls list the mistakes that bias a result without a warning.