How do carousel analytics work?

I asked about this at the community call. It's not high priority and there's no rush, but I'd like to keep the question on our radar.

### Questions

Some more explanation and understanding of seeing carousel analytics data would be fantastic. I see that carousels have a parameter for a unique key (`key=`) which we are now using for both an in-page link anchor and for analytics.

1. How can we access analytics info once a key is assigned?

2. Do analytics keys need to be globally unique?

Anchor tags only need to be unique to the page, but my understanding is that analytics allows us to access info about clicks on each carousel, which makes me think the keys need to be globally unique.

Just in case they are, collection curators who are using the carousel keys as an anchor tag will prefix those tags with a unique collection prefix (such as `lgbtq-poetry` or `august-catfish-month`) so we don't have to go back and change keys in the event the keys do need to be globally unique.

3. What kinds of info does analytics give us?

4. If a curator wants to add analytics to a carousel, is there anything else they should be aware of?

### Issue resolution criteria

When those four questions are answered sufficiently that I can fill in the `key` section of the [Carousel Examples](https://openlibrary.org/help/faq/collections-carousel-examples) howto page.

### Stakeholders

@libjenner @cdrini

Start with the Carousel Examples howto page linked in the issue and investigate the existing carousel analytics documentation or behavior. Answer the four questions about accessing analytics, key uniqueness, available data, and curator considerations, then use those answers to fill in the page's key section.

Written by the indexing model from the issue text.