That said, I don't really need recommendations as to what to try, I just need something that doesn't present me with a zillion clearly crappy options (and this applies to Apple's too.. Stupid-compiled-pamphlet-of-information I'm looking at you). After eliminating those, there's not so much left in the Android Market that I can't go through it in a reasonable amount of time.
But if you want an interesting take on it, how about figuring out which people who are downloading the popular apps first, and tracking their behavior as a cohort. Use the trendsetters to predict the trends.
One way to look at this problem is to create a "Best Of" app list and keep it really clean.
I'll look into that ;)
I haven't quite gotten to building an app for making it easy to recommend apps yet but it's next on my list;)
Another important component to this would be to ask the user which apps they actually recommend as opposed to which ones they downloaded, used once and then just left on their phone.
Right now I am basically taking the crowd sourcing approach by asking people what their top N Favorite apps are. Once I get some more data, I intend to employ some "amazon-style" collective intelligence and create a recommendation engine.
I realize that there are other people working on this problem but to my knowledge no one has solved it yet;)