In my previous post, I wrote about using spectral biclustering for making product recommendations. In this post, I'll build on that marketing example and use the Apriori algorithm for analyzing product purchases This example is, again, taken from Giuseppe Bonaccorso's book Mastering Machine Learning Algorithms.
My roles in data analytics/science so far have always been focused on online marketing applications, such as analyzing the impact of ad campaigns, the engagement of users with a website, or the performance of blog posts. Machine Learning algorithms and Python libraries like scikit-learn can help marketers derive insights from user data and establish connections between their features or actions that otherwise would go unnoticed.
In my search for resources that bridge data science and marketing, I found the chapter Clustering and Unsupervised Models for Marketing, which explains how to use spectral biclustering for making product recommendations. In this post, I share with you my learnings about this algorithm and its implementation in Python, as taken from the book.
Project completed in week 8 (16.11.-20.11.20) of the Data Science Bootcamp at Spiced Academy in Berlin.
This week we learned to make a Markov Chain Monte Carlo (MCMC) simulation of new customers in a supermarket, based on data about customer paths from entrance to checkout through four aisles (fruit, drinks, dairy, and spices) recorded on five days from 7 am to 10 pm. This project was particularly challenging for two reasons: it involved object-oriented programming (OOP) and team work. In this post I'll give you an overview of our workflow.