marketbayesmeta
marketbayesmeta is a focused Python/PyMC library for audit-oriented Bayesian
meta-analysis of marketing measurement evidence. It is designed for small-sample
workflows where evidence comes from geo tests, brand lift studies, MMM, and related
measurement sources.
The core stance is conservative: pool only comparable effects, keep units and uncertainty assumptions explicit, and treat model output as review material rather than automatic reporting.
Current release0.3.0 internal release candidate
AudienceData Science analysts reviewing marketing measurement evidence
ReadinessSupervised analyst use, not unattended production reporting
Where to start
| You want to… | Start here |
|---|---|
| Install the package and run the first workflow | Quick Start |
| Prepare a tracker for modelling | Prepare a Tracker |
| Run a YAML-configured analysis | Run an Analysis |
| Review whether outputs are reportable | Review Outputs |
| Check exact YAML fields | Configuration Reference |
| Check command-line entry points | CLI Reference |
| Understand the Bayesian model | Statistical Model |
| Understand small-sample cautions | FAQ |
Sections
- Tutorials — learning-oriented install and first-run material.
- How-To Guides — task-focused operational guides.
- Reference — exact CLI, config, API, and output artefact reference.
- Explanation — interpretation guidance, release status, and FAQ.
- Methodology — statistical framing for the default model.