Provide example data for testing and documentation

[wkearn]

The example datasets are valuable to have around for testing and documentation, but we should generally avoid committing large binary files to the repository if we can. I think we need a system where we can store example data files outside the repository and download them either when the package is installed or when the user tries to access them.

[Teschl]

I was thinking we could handle example datasets like seaborn. load_dataset() will load an example dataset from the online repository.

This function provides quick access to a small number of example datasets that are useful for documenting seaborn or generating reproducible examples for bug reports. It is not necessary for normal usage.

Using the example DEMs we could add functions equivalent to: seaborn.load_dataset() and seaborn.get_dataset_names(). We still should keep the random generation, so one can observe changes on a smaller, more visible scale. I'll work on getting a pull request ready for this.