First off, congratulations on going live (I'm still not there!), and also on avoiding my pet design peeve: shirt images that are too small. Yours are big enough to see without having to click through to other pages. Well done.
It's good that you are getting positive feedback from people you don't know. That shows that they are somehow finding your site. But it typically takes a lot of exposure to generate sales. Someone may like your design, but not need another shirt, or not want to spend money on one right now. They may be back when it is time to buy a gift for a friend, or themselves--or not. That's why companies spend money to keep their product in front of consumers: to be there when it is time to buy.
How many designs to start out with is a tough call. Each design is an investment of time and money, so one doesn't want to waste that on a design that won't sell, but that is part of the cost, and risk, of starting a business.
You probably have some friends in the target market. If they can be trusted to give honest feedback/opinions and not just say nice things, then you should ask them which of your design ideas they like the best. I sort of did that by accident and ended up selling some shirts simply because a friend liked a design idea so I chose that design when I wanted to test print some shirts. I gave him one, and then other people saw me and him wearing it and wanted it. (Remember, I don't even have a web site up, yet.) So get some people to "vote" on your designs when they are still just ideas or art, and then pay them in T-shirts (and slip them a stack of business cards to hand out when they wear your shirts).
I currently have around 75 design ideas, many of which will probably never end up on a shirt. I'm aiming to launch my site with 18 designs that have been vetted by my reviewers, but that also represent the various "flavors" of my ideas. How well each of those sells will guide which types of designs I add next.
If you can come up with 20 or 25 ideas, maybe you could winnow that down to the best 6 designs and run with those to start, but make sure the range of designs covers the "flavors" that you want to test in the market.
Good luck. Let us know how things progress.