Please, don’t. At least not on social networks like Twitter. Some days ago Eric Schmidt had an interview which is published on ReadWriteWeb and it’s about what web will look like in five years. One of his questions was “We can index real-time info now – but how do we rank it?”.
Google approach to the web is to rank everything. But on Twitter we all are equal to each other, we have the same number of characters to write a message (140, remember?), the number of followers doesn’t matter, time when you joined Twitter is not important, lists or favourites won’t help either. So basically we don’t need a rank system on Twitter.
Dave Winer votes for ranking on his blog. But, what we really need is an exclusion system. First, we need to exclude spam. Then exclude retweets and reposition the original tweet instead of last retweet with the number of retweets for a possible importance level. Favourites could also matter in ranking tweets up, but again, keeping them sorted by date is more important.
After all that being said, he thinks that will be (impossible) possible to rank real time information, but it won’t be that clever.