The frequency response graph can give a general description of what the headphone is like. Being linear without any peaks is a sign that it will sound smooth without any parts of the music being over or underemphasized.

Frequency response is the easiest was to demonstrate how headphones will sound for someone who hasn't heard them. They are some of the only objective sound characterization of headphones. They are used to compare 2 headphones and see why they sound different. It's best to look at data about headphones relatively.

You can EQ the frequency response, but it won't sound the best. A headphone without bass adjusted to sound like a headphone which naturally has a lot of bass won't sound the same. (I'm not saying an EQ will make headphones sound bad)