I have a project and I couldn't understand what I have to do because I am new to retail analytics.

They said "Our goal is to measure the effects of promotion on sales" and "Your goal is to model the effect of promotion on products and stores. In order to answer questions, you should divide products and stores into 3 clusters each. (High, Medium, Low)"

I have two datasets, let's say;

1 - data.csv -> Date, Store, Product, QuantityOfSales

2 - promotion.csv -> PromName, StartDate, EndDate

There are questions like that:

  • What are your criteria for separating High, Medium, and Low Stores? Why?
  • Which items experienced the biggest sale increase during promotions?
  • Are there stores that have higher promotion reactions?
  • What is the biggest effect explaining sales change during promotions?
  • Is there any significant difference between promotion impacts of the High versus Low items?

Also, they gave me the same dataset but different time intervals to measure how well my model has worked on this new data.

Actually, I am stuck because I don't know what exactly I have to do. Can anybody show me the path I should follow or show samples?

  • 1
    $\begingroup$ This is not a great question for SO. It is too general and is primarily opinion based. A better use of SO is to try something and then when you get stuck come back with a more specific question. $\endgroup$
    – MichaelD
    Aug 17, 2019 at 19:31

1 Answer 1


Start by plotting some of your data in various ways in order to get more familiar with it and understand the big picture. For example plot the total amount of sales for every store, for every product, look at the different by store for the same product, difference promotion/not promotion for the same product, etc.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.