I tend to go for the most technically challenging aspects because those have the highest risk of not working as planned or outright failing. You may have to modify everything else to fit these unforeseen changes. On the flip side, you may discover something cool that you want to implement while playing with the editor, but you have to change all your terrain or units to take advantage of it. No point in putting a lot of effort into terrain or units early on if they end up not working with the final version.
I use placeholders until I get the game working properly, then I start adding everything else to take advantage of my strengths and cover up the weaknesses. I also use a ton of dialogs to display data for debugging purposes since you don't really need a UI at this point. If it's an ambitious project then odds are that your final product will be a bit different than what you initially set out to do. One of the hardest things to do is stop improving the map because learning the editor has a snowball effect of opening up more and more cool things that are simply too awesome not to do, even if you have to go back and rework things to accommodate them.