Data Analytics for Marketing

4

Econometrics and Causal Inference with Statsmodels and PyMC

All models are wrong, but some are useful.

– George Box

As a marketing analyst, usually you will not have the luxury of big data to feed into machine learning models. The data will be sparce or made up of low-frequency, time series, or panel data, which will prevent you from brute forcing your way through. You need a solid understanding of econometrics and the principles of causality to answer common questions your stakeholders will have.

As you can recall from Chapter 1, What is Marketing Analytics?, some questions you will encounter in your daily work will be “Why did something happen?”. This is diagnostic analytics. As a marketing data analyst, you will be asked, “What is the impact ...