You want to use the guidelines mentioned in holly's book to pick out "golden" keywords and then ultimately ranking for harder keywords. Even what she states is just a matter of opinion. Maybe, down the line you'll find that 100 searches and 1.5 million is good for a niche as well. You really have to go through the WHOLE process before you know what's a good niche market
When you're thinking about broad and narrow keywords, it's going to be pretty subjective but you WANT narrower(long-tail) keywords as they're more qualified for what the person is searching for. An example would be like art versus bronze metal art classes new jersey. Art could be literally anything so it'd be a poor choice for a niche.
You really can't knit pick too much about the numbers. The numbers give you a feel for the market. Don't spend too much time worrying about keyword research. I know according to one author's ebook he spent 15+ hours nailing down a 'superficial' list of keyword terms. I duplicated the process and it took me just as long and that's just the start...what comes after is useful for what should be on your site but...
If you ask me, this is kinda crazy in the beginning. This might be better later on once you've got your web design down, content and traffic flow to help with SEO but beginners want results fast. So you want eyes on your site ASAP!
Also, the competition doesn't really matter too much if you're an expert or have some kind of 'insider' information in the field. I'm sure there are a bunch of people going for what's hot and have no clue what they're doing then get mad because of dismal results.
You have to think... you're competing in a niche that people know what they're talking about. Go for what you know and work it from there. Once you've nailed this down you can go for what's hot.
BUT WAIT! you might be thinking well I can just learn as I go... why would you do that when you've ALREADY got a head start in what you already know?!
Hopefully this didn't confuse you anymore