To use a module, you have to pick one from the left-side blade and drag it inside the gray area in the middle of the window. That's not all. Each module has an output, which is supposed to either be the final product or to be fed as some other module as input.
Except for a few, the majority of modules also need an input. Once a module is selected, it will be highlighted and you can see whether it requests one or more inputs, along with which kinds of inputs are needed, and you can also see how many outputs it has, along with each kind.
Inputs are represented by little circles on the top of the selected module. Outputs are represented by small circles on the bottom of the selected module. Details for both are shown when the mouse pointer hovers over a circle. Let's call the circles on the top of each module input circles, and the circles on the bottom will be called output circles. The following screenshot shows a module and its input and output-circles:
Many things were done in Figure 13.9. First, the experiment was renamed to R in Cloud. The search bar in the left blade was used to search for Split Data module, which was dragged into the middle zone. Then the module was selected; it requests a single input and can have two outputs. By hovering the mouse pointer over the input circle, it showed the types of inputs accepted.
On the far right, inside the white blade, you can see the parameters related to Split Data module. They are Splitting mode, Fraction of rows in the first output dataset, Random seed, and Stratified split. There is even a small checkbox asking whether the split should be randomized or not.
In the end, all experiments can be described as a bunch of interrelated modules. To connect (relate) one module to another, all a user has to do is to click in the output circle of a module and drag the mouse to the output of the module you wish to connect to. Building data science models with Machine Learning Studio is as easy as that.
Yet, you can make things slightly more complicated and flexible, by deploying raw codes as well. Skills in SQL, Python (both 3.x and 2.x), and R are welcome here. Let's put our knowledge of R and Machine Learning Studio together to build a complete experiment.