Ideas come from brainstorming - sit somewhere with a notepad and a pen and start writing out ideas - something the average person will be interested in, existing ideas than can be improved, something businesses find useful (like consumer feedback) yet have become accustomed to living without.
Creating something entirely new for a need that does not exist is unheard of. Need drives innovation. A niche I would say is something people have been accustomed to not having. Programming like business is either focused on increasing profits(products or features which are money makers) or reducing cost (saving time, money, effort etc.)
If I could, I'd focus primarily on writing something useful that saves time. Trying to build something that's 'too useful' sometimes makes the process slower than it needs to be.
A simple example of the above-mentioned is inventory management systems (compare selling non-perishables vs selling perishables). An electronics store would use a first-in-first-out approach due to tax benefits. A 'food' store would use a first-expire-first-out approach. The latter approach is quite visible when standing in-front of rows of milk with the first to expire at the front and the latest to expire at the back.
An electronic store could identify electronics by type/brand, while foodstuffs requires another level of precision (a batch - all same ingredients used, same expiration date).
Selecting via type is much easier than selecting via type and batch. 'Foodstuffs' that are used as ingredients generally require some level of trace-ability - which adds another step into the process.
This is where pragmatism (business focused) outshines perfectionism (feature focused).
Many times I've found that new ideas on a new heightened level start flowing due to the distraction.