It definitely depends on the audience you are trying to reach. I just finished building a feature selection algorithm to plug into sklearn: https://github.com/chasedehan/BoostARoota

I'm also trying to figure out the best ways to do that. What my plan is: 1) I presented it last week at the ODSC West Conference and have seen a couple people use it and give me some comments. 2) I am planning on incorporating the algorithm into some Kaggle kernels so people can see its value. It definitely works better than some other outlets. 3) Respond on Quora or other places for cases when it might be useful 4) Shamelessly plug on HN ;)