Scikit-learn example for random forest with Joblib

It would be nice to see an example using Scikit-Learn's RandomForest parallelized with Dask using the joblib/dask connection. This is a common question, and so probably has high value.

My understanding is that this will depend on the 0.20 release of scikit-learn first, though I could be mistaken about this.

It would also be useful for this to be smoothly runnable on [dask-examples](https://github.com/dask/dask-examples#contributing)

Start by checking the scikit-learn 0.20 dependency mentioned in the issue and the contribution guidance in dask-examples. The work is done when a runnable example demonstrates a RandomForest parallelized with Dask through the joblib connection and fits the dask-examples contribution requirements.

Written by the indexing model from the issue text.