CI to for Kaggle Docker Image

Kaggle is a popular and rapidly growing hub for data scientists which provides a browser-based ipython notebook for data scientists to work in. We've had a kaggle user attempt to install evalml in this environment built by the docker image, but to no avail. The issue is detailed in #3174, but is connected to #2671 . Ultimately, there are incompatibilities between some of the libraries that are installed via conda on the kaggle environment and using the cell based directive `!pip install evalml`. These incompatibilities prevent users from installing evalml (at least via pip).

As an EvalML dev, I would like CI setup to install evalml into the kaggle docker image environment to ensure that kaggle users have unimpeded access to our library.

Resolution of this issue can be a test added to the github CI to build an environment from the [kaggle docker image](https://github.com/kaggle/docker-python), then either pip install or conda install evalml into that environment. If the proper version of evalml is detected after the environment is built, then the test should pass.