First find a project you find interesting, or otherwise worthy. Then, I wouldn't worry about the # of pull requests, just that they have recently accepted pull requests. Once you know they are open to receiving requests, then look at issues to see what low hanging fruit there may be.