Personally, I'm less interested when the designer-developer/creator aims to include more than one region, and/or every Pokémon in the National Dex.
Thoughts on Regions:
Quite apart from being a lot of work, I believe using multiple regions adds needless complexity to the game, especially if some are locked out so you have to play them in a specific order. Creators also have to consider the question, "how do I balance levels across multiple regions?" Balancing for two is difficult enough - refer to HeartGold & SoulSilver - but three or more would be almost impossible, especially if the creator plans to include Gyms and Elite Four for each region.
My recommendation: Design and develop one region, and put your heart and soul into the maps and the people therein. Show you've put time and effort into the project instead of churning out dozens of subpar maps with subpar NPCs.
Thoughts on Putting Them All In:
I don't think there's any point in jamming all 721 Pokémon (or more, because people will inevitably want to include Gen 7 Pokémon in their fangames) into a single game. Your world does not need every version of, for example, a Pidgey evolutionary line, a Rattata line, and a Caterpie line. Creators also need to account for things like the number of legendary Pokémon and pseudo-legendary evolutionary lines (as minimal as possible, so that when they do show up, it's a rare find, a tough opponent, or possibly important to the storyline), and distribution of Pokémon across the region's routes.
My recommendation: Go through the National Dex, pick out the Pokémon you like. As you read through the Dex, write them down in a Notepad file or something, then total them up at the end. I think about 300 species is a good total to have.
Be sure to note down a good variety of early-game, midgame, and late-game species. Some evolutionary lines are often paired with others: you don't often see Caterpie without Weedle, or Eevee without all of its evolutions.
Side Note, on Unofficial Species:
I chose to include unofficial species in my project (I avoid using the term "Fakemon", as I think it sounds a little degrading), and even then I cut back on my homebrew designs, because there is no sense in creating unofficial duplicates of existing species. Unless, of course, there is a clear difference between the official and the unofficial. For example, the early-game Flying evolutionary line in my own project, rather than being Normal/Flying, is pure Flying.