Customer Analytics in Python
5
Lessons58
Quizzes5
Assignments10
Course description
Customer Analytics in Python is where marketing and data science meet. Data science and marketing are two of the key driving forces that help companies create value and stay on top in today’s fast-paced economy. This course is packed with knowledge, and includes sections on customer and purchase analytics, as well as a deep-learning model, all implemented in Python.A Brief Marketing Introduction
In this course, you’ll get to learn theory, as well as acquire practical skills, applicable to a variety of customer analytics areas. You’ll get to meet your instructors and learn about the topics we’ll discuss during the course. This is the place where we’ll discuss the marketing fundamentals, such as the STP framework and Marketing Mix. They will provide us with motivation as to why we take advantage of certain models in the following sections of the course.
K-means Clustering
We continue with segmentation, this time focusing on a flat clustering technique, namely K-means. We’ll divide our customers into clusters using K-means and determine the customer segments. This is a crucial part of the customer analysis, as it is where we discover the most important characteristics, which define our customer groups.
K-Means Clustering based on Principal Component Analysis
In our final section of customer analytics, we’ll introduce a dimensionality reduction algorithm: Principal Components Analysis (PCA). We’ll combine K-means and PCA to obtain even better clustering results and gain insight about our customers. In addition, we’ll perform the model deployment, which we’ll need in the following sections of our course
Descriptive Analyses by Segments
We’ll analyze our customers by filtering our data by brand and segment, and we’ll visualize the findings. In addition, we’ll compute the revenues by brand and total revenues for each segment. Through the descriptive analysis, we will form our hypotheses about our segments, thus ultimately setting the ground for subsequent modeling.
Modeling Purchase Incidence
We’re ready to create our first prediction model. We’ll use logistic regression and model purchase probability of our clients. We’ll compute the price elasticity of purchase probability and use it to gain insight into our average customers, as well as each of our customer segments. Furthermore, we’ll explore the effects of promotion on our clients’ purchase behavior.
Modeling Brand Choice
Our goal is to understand the brand preferences of our clients. We’ll imagine we work for one of the brands and create marketing strategies for targeting customers to attract them to our own brand. We’ll compute price elasticity of brand choice for our brand, as well as calculate cross price elasticities for a competitor brand to try to stay ahead of the competition.
Modeling Purchase Quantity
We conclude the purchase analytics part of the course with a model for purchase quantity. Here, we’ll use linear regression to determine how many units of our product the customer likes to buy. We’ll examine the effects of promotion on purchase quantity and determine price elasticity of demand.
Deep Learning for Conversion Prediction
Machine learning and artificial intelligence are at the forefront of the data science revolution. We will take advantage of the TensorFlow 2.0 framework to create a feedforward neural network. This is the part where we will build a black-box model, essentially helping us reach 90%+ accuracy in our predictions about the future behavior of our customers.