If you want to finish your thesis, the best advice I can give you is to
select a topic that your advisor is already working on.
If your advisor has already an area of interest, she probably, already has the dataset that you are going to work on. If not you would create that dataset for a future student, in either case your thesis will be part of a much larger picture. Also the importance of the frindship between you and your advisor cannot be overstated.
So in reality, choosing the right advisor is 10 times more important than choosing the right topic.
(I've adivised over 10 thesis, and reviewed several more)
