Towards Data Science 12:23 am on June 3, 2024
The article discusses how to estimate the intrinsic causal influence of marketing campaigns using Causal AI, highlighting its importance beyond traditional models and potential use cases like identifying lagged effects. It explains setting up a graph-based model with observed variables (brand spend, social spend) and noise terms, training it via DoWhy's GCM module to calculate intrinsic causal influence for non-root nodes in a DAG.
- Overview: Estimating marketing campaign impact using intrinsic causal influence.
- Data Generation: Creating a directed graph model representing brand and social spend with noise terms as root nodes.
- Model Training: Utilizing DoWhy's GCM module to train the model and calculate intrinsic causal influence using Shapley value estimates.
- Visualization: Representing results in a bar chart, displaying each node's contribution to the target outcome.
- Practical Implications: Exploring intrinsic causal influence as a tool for effective marketing analysis and potential model improvements.
https://towardsdatascience.com/measuring-the-intrinsic-causal-influence-of-your-marketing-campaigns-aa8354c26b7b
< Previous Story - Next Story >