Recommender Blocks
planned
Noel Tock
We've been looking at various solutions to leverage the Altis Analytics data in order to create machine-learning models for the purpose of recommending content on a user-by-user basis, i.e. personalised content recommendations based on past behaviour and traits.
This is in the backlog, please let us know if you have any thoughts, suggestions, ideas, etc. as we move this into development.
J
Jay Chong
Quick-backs / bounce rate should be in the formula before you can determine the popularity of website journey (from Page A to Page b).
Presuming that you managed to track user (identify unique user) since first time visiting to the website Altis Analytics.
Just my 2 cents.
Robert O'Rourke
Jay Chong: we're collecting an anonymous ID that persists between pages, so the bounce rate is already available in the stats API request, we just don't visualise it yet. We may implement this using AWS Personalise or using ClickHouse's built in ML now that they're adding vector based queries.
Noel Tock
planned