When it comes to your daily life, what are your pain points?
Examine each pain point, document why it's a pain, who else is likely to have that pain, how it might be fixed by software.
Analyze existing software that attempts to address the pain point. Do any of them address it adequately? If so, strike that pain point (you could compete, but it will be harder). If it's addressed, but not adequately - or not addressed at all, then you have a candidate. If not addressed all, think long on whether there is a reason for that, and what it could be.
Interested in software for a particular group (e.g. bicyclists)? Make friends within the group, then find out their pain points, and repeat the above procedure.
Once you have ten 16 candidates (if you can't come up with 16 using this process, then add a new group or switch), pick the best 8 based on the following:
- how executable your idea is
- how crowded the market space is
- your skillset, or the skillsets of the team you are confident you can put together
- how long it might take you to get a product to market
- how much investment is needed, and what your runway would be
Then do the same for that 8 to pick the best 4, and then the best two.
Do a MVP for each of the best two and see what traction you can get.
It's possible all four of your top ones are viable idea.
As for a Twitter alternative, that's like saying an Office alternative. Pick one facet of Twitter (e.g., doing polls) and analyze as above. You're better off with something else though, as the market for Twitter alternatives is already over-saturated. And in the next two-years it will become even more so. Someone will create the next big social media thing though, so who knows.