Convert radviz into FeatureVisualizer about yellowbrick

[rebeccabilbro]

Note that radviz for classifiers and regressors might be different -- might need to do some different things in the way we color them. For classifiers, as many colors as there are classes. Perhaps a gradient for regressors - red-to-blue gradient where grey is the mean, dark red is 2 standards from positive, dark blue is 2 standards from negative.

[bbengfort]

As with parallel coordinates, I'll go ahead and bring the implementation into the yellowbrick code base so that we can optimize it!

[bbengfort]

This was a fun one - I had to dive into Numpy a lot more than I expected (the Pandas implementation relied heavily on DataFrame and Series stuff)!

Same as ParallelCoordinates - this is an initial visualizer, and there is a lot of room for improvement; but this is also where we iterate!

[bbengfort]

Also it doesn't seem fair to close this issue without putting an example in!