Multi touch attribution models for marketing campaigns, including Markov chains. Includes heuristic (ie: first, last, and linear touch attribution) and stochastic (ie; Markov chains attribution) models for attributing customer conversions to marketing channels.
Started by creating a random dataset of customers, channels and conversions. Then, after manipulating the data in the correct format and merging the conversion results for all 4 methods, a final visualization is created to show the "performance" of each model's interpretation. The final visualization is a heatmap of the Markov chain transition probabilities, detailing which transitions attributed the most conversions, which provides insights as to which transitions are valuable (and which aren't).
Next project will include an analysis of a public dataset.