In this dataset, we simulated 768 different building shapes. We made vary the following attributes:

• X1 Relative Compactness
• X2 Surface Area
• X3 Wall Area
• X4 Roof Area
• X5 Overall Height
• X6 Orientation
• X7 Glazing Area
• X8 Glazing Area Distribution

Then we performed energy analysis and recorded two real-valued responses: the heating and cooling load.

Grouping the heating load in 5 classes, we are able to predict what will be your heating load and therefore energy consumption.

Knowing such information allows you to change your design beforehand or perform the necessary renovations if you are not happy with what you see.

In the long run, it's a lot of money and energy saved. Good for you and the environment.