Understanding the metrics intimately, building some intuition for them via dashboards, and plotting things in order to formulate hypotheses is a good way to get you warmed up. It seems that by these projects, they're trying to familiarize you with what they measure.
For projects like improving search relevance being able to formulate an insightful theory and quickly checking this against logs/metrics is way more important than programming skills. It seems that you're working with all the right tools - come up with an insight & make a pitch for an A/B test to the right boss!
My job satisfaction has increased greatly with the ability to drop the ego and take an unglamorous task and just do a good job on it.
Hah, one of my first jobs at Google was to do the dashboards for our feature's dogfood, and my manager had to tell us "I know that engineers love to tweak dashboards and make them look pretty, but please try not to spend too much time on them, we have other important things to do."
It's really common for early versions of dashboards or metrics to use off-the-shelf components because the person who set them up had other responsibilities (like building the product) and didn't have time for it. That doesn't mean they should stay that way forever. They could be giving you this assignment as a test to see what you do with it - if you replace the off-the-shelf stuff with pipelines and displays that are way better, that shows that you have the skills that'll be necessary for working in IR or ML.