2022: Week 9 - Customer Classifications

 Challenge by: Jenny Martin

Recently, I was doing some Tableau Desktop training that went off on a bit of a tangent. After introducing cohort analysis, I was asked whether we could see the year on year difference in the number of customers. This was answered with a quick table calculation. However, the follow up question of how to get a list of the 158 customers who didn't order again in 2019 had me fairly stumped! After lots of playing around and chatting it through with various colleagues, I came up with a solution that used LODs and Parameters, but all the while I was wishing I could solve the problem in Tableau Prep instead. So of course, I did! 

I'm really excited to see the different solutions for this week so don't worry if your instinct is to deviate from the requirements below. Here are the definitions for how we will classify the customers:

  • New = this is the first year the customer has ordered
  • Consistent = the customer ordered this year and last year
  • Sleeping = the customer has ordered in the past, but not this year
  • Returning = the customer did not order last year, but has ordered this year

See full dashboard here

Input

You may recognise this as the sample superstore dataset that comes with Tableau Desktop by default. We're using just the Orders table. 

Requirements

  • Input the data
  • Aggregate the data to the years each customer made an order
  • Calculate the year each customer made their First Purchase
  • Scaffold the dataset so that there is a row for each year after a customers First Purchase, even if they did not make an order
  • Create a field to flag these new rows, making it clear whether a customer placed an order in that year or not
  • Calculate the Year on Year difference in the number of customers from each Cohort in each year
    • Cohort = Year of First Purchase
  • Create a field which flags whether or not a customer placed an order in the previous year
  • Create the Customer Classification using the above definitions
  • Join back to the original input data
    • Ensure that in rows where a customer did not place an order, the majority of the original fields are null. The exceptions to this are the Customer Name and Customer ID fields.
  • Output the data

Output

  • 26 fields
    • Customer Classification
    • YoY Difference
    • Order?
    • Year
    • First Purchase
    • Row ID
    • Order ID
    • Order Date
    • Ship Date
    • Ship Mode
    • Customer Name
    • Segment
    • Country/Region
    • City
    • State
    • Postal Code
    • Region
    • Product ID
    • Category
    • Sub-Category
    • Product Name
    • Sales
    • Quantity
    • Discount
    • Profit
    • Customer ID
  • 10,396 rows (10,397 including headers)
You can download the full output here. 

After you finish the challenge make sure to fill in the participation tracker, then share your solution on Twitter using #PreppinData and tagging @Datajedininja, @JennyMartinDS14 & @TomProwse1

You can also post your solution on the Tableau Forum where we have a Preppin' Data community page. Post your solutions and ask questions if you need any help! 


Popular posts from this blog

2023: Week 1 The Data Source Bank

Image
 Created by: Carl Allchin Welcome to a New Year of Preppin' Data. These are weekly exercises to help you learn and develop data preparation skills. We publish the challenges on a Wednesday and share a solution the following Tuesday. You can take the challenges whenever you want and we love to see your solutions. With data preparation, there is never just one way to complete the tasks so sharing your solutions will help others learn too. Share on Twitter, LinkedIn, the Tableau Forums or wherever you want to too. Tag Jenny Martin, Tom Prowse or myself or just use the #PreppinData to share your solutions.  The challenges are designed for learning Tableau Prep but we have a broad community who complete the challenges in R, Python, SQL, DBT, EasyMorph and many other tools. We love seeing people learn new tools so feel free to use whatever tools you want to complete the challenges.  A New Year means we start afresh so January's challenges will be focused on beginners. We will use dif
Read more

2023: Week 2 - International Bank Account Numbers

Image
Challenge By: Jenny Martin For week 2 of our beginner month, Data Source Bank has a requirement to construct International Bank Account Numbers (IBANs), even for Transactions taking place in the UK. We have all the information in separate fields, we just need to put it altogether in the following order: Inputs A list of the transactions, with information about the receiving bank account  A lookup table for the SWIFT Bank Codes  Requirements Input the data In the Transactions table, there is a Sort Code field which contains dashes. We need to remove these so just have a 6 digit string ( hint ) Use the SWIFT Bank Code lookup table to bring in additional information about the SWIFT code and Check Digits of the receiving bank account ( hint ) Add a field for the Country Code ( hint ) Hint: all these transactions take place in the UK so the Country Code should be GB Create the IBAN as above ( hint ) Hint: watch out for trying to combine sting fields with numeric fields - check data types Re
Read more

How to...Handle Free Text

Image
Free isn't always a good thing. In data, Free text is the example to state when proving that statements correct. However, lots of benefit can be gained from understanding data that has been entered in Free Text fields. What do we mean by Free Text? Free Text is the string based data that comes from allowing people to type answers in to systems and forms. The resulting data is normally stored within one column, with one answer per cell. As Free Text means the answer could be anything, this is what you get - absolutely anything. From expletives to slang, the words you will find in the data may be a challenge to interpret but the text is the closest way to collect the voice of your customer / employee. The Free Text field is likely to contain long, rambling sentences that can simply be analysed. If you count these fields, you are likely to have one of each entry each. Therefore, simply counting the entries will not provide anything meaningful to your analysis. The value is in
Read more