Is there a relation between the number of times a paper is downloaded and the number of citations?