There's a simple test I use for these cases... it's called the "Least Restrictive Subset Test", or LRST for short. You apply it to see if targeting a specific group is worthwhile.

It works by asking two simple questions:

1) If I marketed to a less restrictive subset, would I capture a smaller percentage of my target audience (in proportion to the population)? For example, if I said, "GirlsCode" -- would I still get ~10% black females attending? I suspect you'd get the expected percentage of white females attending, so if white girls were my target audience, I needn't restrict the audience any further. But for black females I would.

2) If I market to this restricted subset will I likely increase the total number of my target audience in attendance. For example, if I said, "WhiteMenCode" do I expect more white men to attend than if I said "MenCode"? I suspect probably not (at least not people serious about coding, versus just trying to prove a point about race). But BlackGirlsCode probably will increase the total number of black females who are serious about learning to code.

I think BlackGirlsCode passes the LRST.