Dython

* Added `nominal.identify_columns_with_na` (by **[musketeer191](https://github.com/musketeer191)**)
* Added `nominal.identify_numeric_columns` (issue [58](https://github.com/shakedzy/dython/issues/58), by **[musketeer191](https://github.com/musketeer191)**)
* Added `nominal.identify_columns_by_type`
* `nominal.identify_nominal_columns` no longer accepts the `include` parameter (use `nominal.identify_columns_by_type` instead)
* Fix docstring of `nominal.compute_associations` (issue [55](https://github.com/shakedzy/dython/issues/55))
* Requires Pandas 0.23.4 or greater (was required before, but not specified in setup file)

* Resolve issues [48](https://github.com/shakedzy/dython/issues/48) and [#49](https://github.com/shakedzy/dython/issues/49)

* Fix issues [28](https://github.com/shakedzy/dython/issues/28), [#31](https://github.com/shakedzy/dython/issues/31), [#41](https://github.com/shakedzy/dython/issues/41), [#46](https://github.com/shakedzy/dython/issues/46)
* `nominal.cramers_v` can be used without bias correction
* Removed `kwargs` from all methods, replaced with explicit API
* `nominal.associations` and `model_utils.roc_graph` now return a dictionary of output values
* `model_utils.roc_graph` can accept an `ax`
* license replaced to BSD-3

* `nominal.associations` now handles single-value features (issue [38](https://github.com/shakedzy/dython/issues/38))

* Added log-base selection in `nominal.conditional_entropy` (issue [35](https://github.com/shakedzy/dython/issues/35), by **[ahmedsalhin](https://github.com/ahmedsalhin)**)
* Added new example: `associations_mushrooms_example`
* Renamed example: `associations_example` is now `associations_iris_example`

* Requires Python 3.5+
* Private methods and attributes renamed
* Fixed incorrect `__version__` varaible