In this use case we are going to embed our user data in order to draw relevant conclusions our user acquisition. Recently we have run 2 campaings - an older one at the end of August, and a more recent one right before Christmas, 2023. The first campaign operated with more generic messages, while the latter one was aided by XYZCr$$d, a made up influencer. In terms of user data, we have
  1. the ad creative the users clicked on
  2. the signup date as a unix timestamp
  3. and the average daily activity (measured in API calls / day)
Embedding those inputs in a vectorspace allows us to cluster them and find meaningful user groups. To understand the embeddings we will
  • create a UMAP visualisation
  • and examine the cluster labels’ association with initial features

​
Follow along in this Colab

Analytics - User Acquisition Notebook

Open the interactive Colab notebook to follow along with the user acquisition analytics tutorial.