Break scales when the data has outliers

[JackWilb]

Add a bin for "larger than "

[JackWilb]

Potentially useful library https://github.com/lukewhyte/d3-scale-break

Seems like I'll just calculate IQR, and then make bins for high and low outliers based on the box and whisker plot algorithm. That should be robust enough for skewed data.

I'll also need to add the ability to switch the scale type (to log, for example), which will then need to use this code