4 Using generative AI for result interpretations

 

This chapter covers

  • Building analysis context for generative AI by building an analysis summary
  • Ways to describe visual results to a language model
  • Using generative AI’s advice to connect analyses
  • Interrogating generative AI for business conclusions of analysis results

In the previous chapter, you performed a number of descriptive analyses and ended up with quite a few charts and some statistical modeling parameters. Yay! You can now bag it all and send it to your boss, right? Maybe . . . Depends on your boss, really.

In your work, you may encounter different types of managers. Some will be really well-versed in analytics themselves, and you’ll be cooperating with them, learning a ton of useful stuff. They will accept your results and discuss assumptions and methods with you; they may ask you to do some analytics neither you nor your generative AI advisor expected in this context. All will be good. The kind of boss you need to be wary of is the type with an overinflated opinion about their analytical skills. We won’t dwell on this type, as dealing with these people is a matter of psychology, rather than data analytics.

4.1 Problem definition

4.2 Popularity of product categories

4.3 Performance of products in their categories and regions

4.4 Review scores distribution analysis

4.5 Order status

4.6 Relationship between product attributes and the shipping costs

4.7 Relationship between product, transaction, shipping attributes, and the review score

4.8 Differences in sales performance and customer satisfaction between sellers

Summary