Good morning Jober-
Sorry to hear of the demise of the Mazda... But it seems to have kick-started your Locost ambitions, so maybe some good will come of it. (Or some really fun "Eee-ville"!!!
There's a valid argument for both sides of your coin, I think. (Uh-oh, here he goes again...
Side one is, take what you've got, in this case a FWD Mazda drivetrain that's paid for, and build around it. That's a big part of the "Locost Philosophy" after all. You can figure out how to build a middie using those components while you practice your welding.
Side Two: It's easier to do a more basic 7-clone type of Locost, especially with a Miata donor. Lots of previous builds to draw from, lots of support for the Miata components as far as hot-rodding or just replacement bits. Conventional wisdom would dictate that this is the easiest type of build to finish, because it's such a well known path.
Sleep on the build type, think about it. Listen to the voices in your head. (If you don't have any now, you will, just start building!) If I were in your place, I think I'd go with the middie, but that is hardly a recommendation. (Nobody ever said I was bright...) It just seems like it would be more fun to use what you got in some unique fashion, totally your concept and (mostly) your design. While not every build in here will be relevant, the concepts used in building a "standard" chassis as far as material size/type and triangulation are the same. The design of the front suspension is the same, no matter where you put the engine, same is true of brakes and steering. And you're halfway there, got the engine/trans/axles/hubs/brakes/etc sitting in the garage.
Either way, I think your next step is to get a welder and do some practicing. Keep us posted on the build, I know I'll be waiting to see what comes of it...