Here is the current iteration of my OK Cupid portfolio project:
I keep coming back to this data set for a couple of reasons. For one, as I learn more things I realize I can make this project better. And for another, it is such a messy, “real,” complicated data set! Kudos to the developers for not “prettying” up the data and letting us have it as they got it. I won’t lie, this project frustrated me a lot … but I also learned a lot by doing it.
Now that I’m just about done with the deep learning skill path, I’ll be revisiting this project yet again but this time applying deep learning models to the data. It’s gratifying to see different machine learning models converge to about the same place in terms of accuracy and f-1 score.