Customer Analytics in Python

https://365datascience.teachable.com/p/customer-analytics-in-python

A Brief Marketing Introduction

• Start Segmentation, Targeting, Positioning
• Start Marketing Mix
• Start Physical and Online Retailers: Similarities and Differences.
• Start Price Elasticity

Segmentation Data

• Start Getting to know the Segmentation Dataset
• Start Importing and Exploring Segmentation Data
• Start Standardizing Segmentation Data

Hierarchical Clustering

• Start Hierarchical Clustering: Background
• Start Hierarchical Clustering: Implementation and Results

K-Means Clustering

• Start K-Means Clustering: Background
• Start K-Means Clustering: Application
• Start K-Means Clustering: Results

K-Means Clustering based on Principal Component Analysis

• Start Principal Component Analysis: Background
• Start Principal Component Analysis: Application
• Start Principal Component Analysis: Homework
• Start Principal Component Analysis: Results
• Start K-Means Clustering with Principal Components: Application
• Start K-Means Clustering with Principal Components: Results
• Start K-Means Clustering with Principal Components: Homework
• Start Saving the Models

Purchase Data

• Start Purchase Analytics - Introduction
• Start Getting to know the Purchase Dataset
• Start Importing and Exploring Purchase Data
• Start Applying the Segmentation Model

Descriptive Analyses by Segments

• Start Purchase Analytics Descriptive Statistics: Segment Proportions
• Start Purchase Analytics Descriptive Statistics: Purchase occasion and purchase Incidence
• Start Purchase Analytics Descriptive Statistics: Homework
• Start Brand Choice
• Start Dissecting the revenue by segment

Modeling Purchase Incidence

• Start Purchase Incidence Models. The Model: Binomial Logistic Regression
• Start Prepare the Dataset for Logistic Regression
• Start Model Estimation
• Start Calculating Price Elasticity of Purchase Probability
• Start Price Elasticity of Purchase Probability: Results
• Start Purchase Probability by Segments
• Start Purchase Probability by Segments - Homework
• Start Purchase Probability Model with Promotion
• Start Calculating Price Elasticities with Promotion
• Start Calculating Prcie Elasticities without Promotion: Homework
• Start Comparing Price Elasticities with and without Promotion

Modeling Brand Choice

• Start Brand Choice Models. The Model: Multinomial Logistic Regression
• Start Prepare Data and Fit the Model
• Start Interpreting the Coefficients
• Start Own Price Brand Choice Elasticity
• Start Cross Price Brand Choice Elasticity
• Start Own and Cross-Price Elasticity by Segment
• Start Own and Cross-Price Elasticity by Segment: Homework
• Start Own and Cross-Price Elasticity by Segment - Comparison
• Start Brand Choice Models: Homework

Modeling Purchase Quantity

• Start Purchase Quantity Models. The Model: Linear Regression
• Start Preparing the Data and Fitting the Model
• Start Calculating Price Elasticity of Purchase Quantity
• Start Calculating Price Elasticity of Purchase Quantity: Homework
• Start Price Elasticity of Purchase Quantity: Results
• Start Improving the Model: Homework

Deep Learning

• Start Introduction to Deep Learning for Customer Analytics
• Start Exploring the Dataset
• Start How Are We Going to Tackle the Business Case
• Start Balancing the Dataset
• Start Preprocessing the Data for Deep Learning
• Start Outlining the Deep Learning Model
• Start Training the Deep Learning Model
• Start Testing the Model
• Start Obtaining the Probability of a Customer to Convert
• Start Saving the Model and Preparing for Deployment
• Start Predicting on New Data

https://www.toptal.com/r/social-network-analysis-in-r-gephi-tutorial

What are sociograms?

Sociograms are graphs in which each node is a person and the edges represent interactions between them.

Why are sociograms important?

Sociograms are important because they give us an abstraction (and illustration) of how people interact in huge groups. They are simple but quite meaningful. They help us understand societies from many perspectives, such as user centrality, information spread, community identification, and more.

Segmentation

A useful segmentation should include these six characteristics:

1. Identifiable

   You should be able to identify customers in each segment and measure their characteristics, like demographics or usage behavior.

2. Substantial

   It's usually not cost-effective to target small segments - a segment, therefore, must be large enough to be potentially profitable.

3. Accessible

   It sounds obvious, but your company should be able to reach its segments via communication and distribution channels. When it comes to young people, for example, your company should have access to Twitter and Tumblr and know how to use them authentically - or, as Clearblue smartly did, reach out to celebrities with active Twitter presences to do some of yourmarketing for you.

4. Stable

   In order for a marketing effort to be successful, a segment should be stable enough for a long enough period of time to be marketed to strategically. For example, lifestyle is often used as a way to segment. But research has found that, internationally, lifestyle is dynamic and constantly evolving. Thus, segmenting based on that variable globally might not be wise.

5. Differentiable

   The people (or organizations, in B2B marketing) in a segment should have similar needs that are clearly different from the needs of other people in other segments.

6. Actionable

   You have to be able to provide products or services to your segments. One U.S. insurance company, for example, spent a lot of time and money identifying a segment, only to discover that it couldn't find any customers for its insurance product in that segment, nor was the organization able to design any actions to target them.

https://towardsdatascience.com/market-mix-modeling-mmm-101-3d094df976f9

https://www.toptal.com/r/social-network-analysis-in-r-gephi-2

Intro