I would start by considering my motivation. It seems that it is important for you to do something new, rather than re-creating something that has already been done. Is that because you want to be self-affirmed as an innovator? Or gain approval from colleagues?
Or is it because you want to sell your work and you think you have less chance if others have already done it?
Figure out what your motives are. Then decide whether those are the motives you really want. Once you've confirmed that your motives are 'pure you', set your goals for the projects to be in line with your motives.
To continue with the examples above, if you want affirmation as an innovator, spend much more time determining how innovative your idea is before diving into it. Include talking to developer friends and knowledgeable persons you know.
If you want to sell your work, figure out what is out there already, what people are charging for it and importantly, how much demand is there? If you find it is a crowded market, it could be that a) there is a large demand for it and people pay well for it or b) it is a simple thing to build and an obvious idea, so a lot of devs build it.
If you mainly want to grow your skill set, look for work that is already out there and which you find intriguing. Use the project to find out how it works.