The hardest part is to know what actually people _want_ because many product developer make things which they assume people want while in reality they don't.
There's no such resource available to learn what people are actually looking. Atleast not an easy way.
I don't agree that's the hardest problem. Most people want something that solves a problem, and if you see a problem that lots of people have it's not too hard to think up any number of solutions. The really hard part is finding a solution that most of those people think is better than the problem. So many solutions simply push a problem from one area to another (time to set up, cost, annoyances, etc.)
Or start with a hypothesis based on your past personal or professional experience.
The point is that you should try to validate consumer demand as quickly as possible without wasting too much time or money. It's the "Lean Startup" approach:
For example, I recently tried to rent a house in London. The process was painful and time-consuming and I think I could improve it by writing a great software package. But instead of spending 2 years and several million dollars building that software package to a "vision", I would put up a simple landing page, market it to potential customers and perform the main parts of the service manually to gauge consumer interest.
Tom it's no so common otherwise most of the problems would have solved by now. Not everyone gets a "personal itch" and those who even get, there's no guarantee other share the pain. One just can't "get out of the building" and ask people,"What's your problem? I am your Genie!" Most of the time you can't solve a problem because it's out of your domain.
There's no such resource available to learn what people are actually looking. Atleast not an easy way.