Data science is all about clever people wielding magic wands and coming up with fancy models that no one can understand. Or is it?
Take a decision tree. Imagine you are fast asleep, your alarm clock is ringing and you need to decide if you hit snooze or if you actually want to get up. You have two parameters, your tiredness level on an imaginary scale of 0-100, and time of day. If your tiredness level is above a certain threshold, say 80, you will hit snooze. If it’s late in the morning and you know you’ll be late for work if you don’t get up pronto, you might decide to get up. And then you will have all the in-between levels of tiredness and lateness that you will have to consider.
That’s essentially how a decision tree works – you take a couple of features (like tiredness level and time of day), and depending on these the tree will walk through a couple of steps (Am I so tired my brain is not yet functioning? Will I be late for work if I don’t get up?) till it comes up with a decision (hit snooze vs. get up).
And now you are a data scientist. Congratulations!
(And if you want to get your hands dirty and actually work with decision trees go check out scikit-learn. The world’s your oyster.)