Starbucks-Customer-Clusters

Exploratory data analysis of Starbucks customers and grouping them into clusters.

You can see my blogpost for this project here: https://medium.com/@paytonsoicher/what-type-of-starbucks-customer-are-you-72f70e45f65

Project Overview

The Starbucks project was my capstone project for the Udacity Data Scientist Nanodegree. This data was provided by Starbucks to simulate their customers and transactions to see if there are better approaches to sending customers specific promotional deals.

My motivation for this project is that I really enjoy unorthodox projects and I had never done any type of clustering project with scenario data before, so I really wanted to get out of my comfort zone for this one. I also enjoy finding information about businesses that would typically not be viewed as important and taking more of a statistical approach to solving a problem that wouldn't use traditional statistics. I love problem solving, and not having any definitive end result gave me the freedom to solve a problem any way that I could find out.

The different files that were used to create this project are located in the Profile Portfolio and Transcript Data.zip file. Here is a quick overview of each file:

Profile - Dimensional data about each person, including their age, salary, and gender. There is one unique customer for each record.
Portfolio - Information about the promotional offers that are possible to receive, and basic information about each one including the promotional type, duration of promotion, reward, and how the promotion was distributed to customers
Transcript - Records show the different steps of promotional offers that a customer received. The different values of receiving a promotion are receiving, viewing, and completing. You also see the different transactions that a person made in the time since he became a customer. With all records, you see the day that they intereacted with Starbucks and the amount that it is worth.

One important segment of any business is to look for areas that can be improved to save money or create more revenue. I was determined to look to see how well certain promotions were being used by all customers and to try and see what was a more effective method to offering promotions. Once I could see how people normally reacted, the problem then became to find a better way to make sure that the right people were receiving the right promotions.

I used exploratory analysis to see what type of customers existed in the Starbucks universe, their transaction / promotional engagement habits, and created a customer matrix dataframe that gae the whole view of what type of customer they were. The best machine learning approach was then to use a clustering method that could break up customers into different segments that would respond differently to different promotional deals.

By doing this, Starbucks can allocate their promotional deals in a more effective manner. People who come back on a regular basis dont need buy one get one free deals, but should receive discount deals. People who might not be able to afford coming into Starbucks stores at a constant basis or don't go regularly for their own reasons should be given BOGO deals to reengage their interests in Starbucks and make them a more reliable customer.

My results showed that Starbucks customers fall into 4 different categories based on mostly their income and the number of times they interact with Starbucks. There are 2 clusters that come consistantly to Starbucks that don't need to receive buy one get one free offers and should only receive discounted promotions. That way those customers still feel rewarded for being a customer, while Starbucks saves some money that they were giving away to them when they didn't have to. The other 2 clusters are groups of people who should be offered more buy one get one free promotions. These customers might not be able to afford coming into Starbucks stores at a constant basis or don't go regularly for their own reasons should be given BOGO deals to reengage their interests in Starbucks and make them a more reliable customer.

Profile Portfolio and Transcript Data.zip

Three JSON files that show profiles of customers, promotional deals that are offered, and the transaction history of customers.

portfolio.json

id (string) - offer id
offer_type (string) - type of offer ie BOGO, discount, informational
difficulty (int) - minimum required spend to complete an offer
reward (int) - reward given for completing an offer
duration (int) - UNKNOWN
channels (list of strings)

profile.json

age (int) - age of the customer
became_member_on (int) - date when customer created an app account
gender (str) - gender of the customer (note some entries contain 'O' for other rather than M or F)
id (str) - customer id
income (float) - customer's income

transcript.json

event (str) - record description (ie transaction, offer received, offer viewed, etc.)
person (str) - customer id
time (int) - time in hours. The data begins at time t=0
value - (dict of strings) - either an offer id or transaction amount depending on the record

Starbucks Capstone Project.ipynb

Jupyter Notebook that shows the start to finish of the this project. This starts with data cleanup and exploration, top level statsitics of customers and their transactions, and cluster analysis for better approaches of sending customers promotions.

anchorp34 / starbucks-customer-clusters Goto Github PK