Why should I believe such a self-assertion, especially when the claimed attribute in question is generally considered a 'good thing'?
It's easy to claim eclectic musical taste, just as it's easy to claim a wide & varied range of reading material (I'm flipping back and forth between Proust & Homer's Iliad as I write this). Both claims are like a personal profile saying 'attractive' - sure, sure, I believe you but show me the head shot anyways.
Now, if the claim for eclectic were supported by demonstrated variety in listening habits, that might be a different situation. For instance, if the user's play list showed they listened to Bjork's Greatest Hits, followed by Debussy, I might start to believe that their taste was indeed eclectic. Bad, but eclectic nevertheless.
Seems to me that this sort of attention data (for which the effort of spoofing would be greater than any value derived therein) is therefore somewhere between self-asserted & 3rd-party asserted identity in terms of it's 'believability' (all else being equal).