Feature Request: Attribute data clustering

This may be overly complicated, but it would be great if Tippecanoe could calculate a attribute data cluster ([Jenks](https://en.wikipedia.org/wiki/Jenks_natural_breaks_optimization), specifically) for use in chloropeth styling on tilesets, much in the same way as `--calculate-feature-density` allows for styling of heatmaps/point layers.

The way I currently style these layers is by doing equal breaks between the min and the max values output in the metadata, but this is problematic for tilesets with outliers.

Note: I currently use this library for calculating attribute data clustering of geoJSON layers ([https://github.com/simogeo/geostats](https://github.com/simogeo/geostats))

Start by reviewing Tippecanoe's existing --calculate-feature-density behavior and how attribute ranges are exposed in tileset metadata. Define how Jenks natural breaks should be calculated for feature attributes and how the resulting clusters would be consumed for choropleth styling. Done should include an agreed output and behavior for outliers, but the issue does not identify implementation files or tests.

Written by the indexing model from the issue text.