Think of it as an interactive notebook where you can not only write down notes, but also enter data and code and interact with it. You can then draw nice graphs of the data and explore it interactively.

The nice thing about it is that it is all repeatable and that you can share it with other people/researchers.

A good demo would be to watch https://www.youtube.com/watch?v=lO7LhVZrNwA. This was a real time demo at Spark Summit last year.

Note: The demo gets pretty interesting at https://youtu.be/lO7LhVZrNwA?t=36m50s. The presenter handles the situation very well afterwards !

Notebook in this example is much like the notebook concept in ipython.

Basically, it is a literate program that allows you to intersperse code, documentation and visualizations in the same "file" called a notebook.

If you are familiar with how mathematica works, it's basically the same thing.

If you do data-science at all, or even get tired of working in excel, I recommend picking up ipython and matplotlib.