My brother triggered me about making a "City Advisor". He said he don't know of any good site that does this service. And I think he's right. I couldn't find any!
And two of them ask for you to start by tipping an address. That's nice if you already know where you might want to go regardless of anything, but it makes no sense if you're looking for good advice on cities and you don't know about every possibility in the world!
Looks like there's a huge niche there and I'm currently wondering why nobody really tackled it.
I also wonder how much more of a generic system this could be...
Instead of recommending cities only, it could be a system to recommend anything! (What triggered me into this was getting refused by a handpicked process yet again, such as country visas.)
The tricky part is that any recommendation is based upon people and you can't advise upwardly everything, most will have to be left out. A product is made by people, if you say it's bad, you're also saying those people made a bad job there. A city is made by many people, if you say it's a bad place to live, you're also saying "everyone there sucks". Ultimately this could advise even about people themselves!
Personalization. "Good" and "bad" are highly relativistic aspects and it depends on who you are. So we'd need to create a huge profile for each user and customize the experience for them. Then we can start recommending based on that.
And how you begin building something like this?
Like anything else. Making a minimum viable project. First find a team interested in the idea, then see what they can offer to contribute in making it real, finally make something visible in less than 1 week, ideally 1 weekend. It doesn't have to work, it has to clearly communicate the idea to a 8 year old.
This project could start as a City Advisor. It should probably start as one specific item advisor... But keeping in mind the ultimate much bigger goal: building an AI capable of identifying an user and profiling them for whatever we're going to recommend 'em.