As this question is currently worded, it's not a good fit for the site. However, the question of "What factors should one consider when building a cube" isn't a terrible one, and it's one you need to step through in order to answer your question anyway, so that's the angle I'll approach it from.
If you're going to make a cube with a theme, you should consider adding some theme-hate. So you have three ways to get cards from the graveyard to the battlefield. Do you have ways to stop people from doing that? Cards that empty graveyards (with exile) are a great fit. They fit in nicely with your theme in that situational cards are now incredibly relevant, and allows you to have an answer for very powerful cards. Remember that one of the defining characteristics of Magic from its inception is the notion that if you're still playing, you can still win: coming back with an awesome bomb completely changes the board position. (Realistically, that isn't always true... but it is true that every element that enters a set has a counter to it. Unless development screws up and makes another Jace or something...)
As for how many cards, that is a very tricky question to answer. Let's consider the population: you have 540 cards. How many of each color are you going to have? Maybe about 100 of each and 40 colorless? If that's the case, how many blue, white, red, and green reanimator are you going to have? Are there even 500 reanimator-themed cards you'd want to play? I'm not convinced there are. I suspect you're going to need more along the lines of 15 distinct themes (that's still 30 cards per theme) in addition to some vanilla or non-theme creatures.
Bottom line is, building a good cube is not easy. Consider the difference between building a standard deck (one very focused theme, 36 spells, 10-12 unique spells) and a commander deck (one loose theme, 4 or 5 mechanics that synergize well but were originally unrelated, 60 spells, 60 unique spells). Well now you're building a 5 color, 30 theme deck with between 540 unique spells!! By my estimation, about 10 times more difficult than a good commander deck. And if you've ever played commander, you know the first time you play a commander deck, you go "What was I thinking? This card sucks in this deck!" How much complex will the playtesting be to make a good cube?
I'm not saying this to discourage you, but just to prepare you for the task at hand. Even leaning more toward one color, I still wouldn't suggest having more than 25% of the cube be a single color, and that can make "single-theming" very difficult.