Working with datasets

In the previous sections, you were configuring compute and datastore resources under the Manage section of the studio. With this infrastructure configured, you can start pulling data into your registered datastores and register datasets in the Assets section of the studio:

Figure 4.41 – Datasets in the Assets section of the Azure ML Studio experience

Datasets is an abstraction layer on top of the data that you are using for training and inference. It contains a reference to the physical data's location and provides a series of metadata that can help you understand their shape and statistical properties. When you want to access the dataset, you can reference it via its name, and you don't have to worry about credentials or exact file paths. Moreover, all the data scientists working on the same workspace can access the same datasets, allowing them to experiment on the same data in parallel.

There are two types of datasets...