First, let's define the names of the variables that we are going to use in this analysis. We have a list of the numerical variables and a list of the categorical variables. Then, we will redefine our housing DataFrame with a DataFrame that contains only the variables that we just defined. Then, we use the shape attribute to see the size of the new DataFrame:
![](https://static.packt-cdn.com/products/9781789531701/graphics/assets/1b7fa418-b55d-4eb2-887f-2569b754e110.png)
In the preceding diagram, we can see that the shape of our DataFrame has now changed because we have only 15 columns.