Once the model's status transitions to "success," the Causal Graph will be accessible for users to view in the Insights tab.
Causal Discovery Graph
The Causal Discovery Graph is depicted as a Directed Acyclic Graph (DAG), where each node represents a variable, and each directed edge indicates a causal relationship between variables. This graphical representation aids in understanding the causal dynamics within marketing datasets, facilitating the identification of variables that significantly impact marketing outcomes.
Average Treatment Effect (ATE)
ATE quantifies the average influence of altering one variable on another across the entire dataset. It is crucial to evaluate the general effectiveness of marketing strategies.
The Causal Discovery Graph incorporates filtering options to refine the visualization based on the significance and confidence of causal relationships:
Edge Weight Threshold: Utilizes ATE values to filter causal relationships based on their impact. A slider allows for adjustment from minimum to maximum ATE values, facilitating the focus on significant relationships.
Edge Confidence Threshold: Determines the reliability of causal predictions. Setting a threshold filters out predictions below a certain confidence level, ensuring decisions are based on robust insights.