Marketing mix modelling
Adstock, saturation, holdout validation, Bayesian posterior intervals, contribution, and ROI, with the methodology written down next to the code.
Marketing Effectiveness Lab
An open reference implementation for marketing measurement: MMM, connector validation, incrementality evidence, CRM experimentation, learning records, and profit-aware budget planning.
What it covers
Causal measurement
Geo-lift and conversion-lift evidence, quality scoring, and experiment-calibrated contribution reconciled against the model.
Decisions under uncertainty
Profit-aware budget optimisation, uncertainty intervals, and readiness gates a number has to clear before it reaches a stakeholder.
Scope and boundaries
What it is
An open, Apache-2.0 reference implementation: a reusable Python package with a Streamlit dashboard, run on demo and real public data.
What it is not
Not a commercial product. There is no pricing or packaging, and no authentication or governed storage for confidential company data.
Where the work continues
New analytics and measurement work continues in Responsible Neobank Growth. The historical roadmap records the extensions considered when this repository was archived.
Stack
Python
Streamlit
Pandas
Statsmodels
Bayesian regression
Budget optimisation
Connector templates
Weekly assembly
Source diagnostics
Learning library
Historical experiment-registry direction
Historical audit direction
Plotly
Pytest
uv
GitHub Pages