So if we check out the data types of our features, one of the methods we can use the built-in pandas. Okay, so looking at our data, using pandas, the first five rows, we've got values over in our avg_resting_heartrate, the avg_active_heartrate, our BMI, our VO2, the number of workout minutes, numbers of steps, and then the lifestyle. If we're working with much larger datasets, of course we want to spend that across a cluster.

We're going to convert this sort of pandas again because our dataset is small enough that we can fit this all on a single node. We don't have any use case now where we need missing values, so we're going to leave that alone. We don't need to do that now in this case. In the last demo, we talked about how we injected missing values. So let's see what that table looks like, again, it should be the same as the previous one. them all together and now we have our final table. And then we're going to take the average of several different columns. I'm going to grab our user metrics lifestyle column.

So we're going to run the same SQL command as we did in the last notebook. Okay, so after our Classroom-Setup has been run, we're going to set up our user level table. The first thing to do is run our run Includes/Classroom-Setup which has our data fetch and also number different helper functions.

We're going to talk through a number of different exercises that we take in a categorical feature and then transform it into numerical one, so maybe like downstream we could consume this with a machine learning model. And we're going to talk about how we use one hot encoding to represent categorical features numerically. All right, great, in this demo we're going to open up lesson 3.1.2, One-Hot Encoding notebook.

One hot encoding in sql