Application of the algorithm to a dataset containing several features

Dear all,
I just finished the specialization, but a very important question, relevant for the application to my work, persists. In the whole course, we process a single time series, which would be, the response, for instance. With effect, one of the dimensions on the first layer is always 1. I was wondering how could I expand it to a dataset containing several predictive variables. For example, we split the time series Y and fed the neural networks with this splitted values, whereas part of them was reserved for validation.

But, in case I wanted to train the neural networks with a more complex dataset containing several predictors X for the response Y, how should I split the data and fed it into the NN, and which modifications on the declared shape would be needed. It would be great if an example was available, because it is not perfectly clear to me.

Please see this link.