Reference Issues/PRs

What does this implement/fix? Explain your changes.

This PR introduces a new feature, which is an extension of the Individual Conditional Expectation (ICE) Plots.
So far a single color for the ICE lines could be chosen by the user (in this case blue):

display = PartialDependenceDisplay.from_estimator(
    hgbdt_model,
    X_train,
    **features_info,
    ax=ax,
    **common_params,
)

This extension allows coloring each line according to the values of another feature. For this, the user just has to pass a list of values. In this case, we use the discrete "workingday" feature from the bike rental dataset. True indicates workingday, false the weekend. This is as easy as just passing the series as the "color" parameter in the ice_lines_kw:

display = PartialDependenceDisplay.from_estimator(
    hgbdt_model,
    X_train,
    **features_info,
    ax=ax,
    **common_params,
    ice_lines_kw={"color": X_train["workingday"], "palette": "Paired_r"},
)

It also works with continuous features, for example, the hour of the bike rental:

display = PartialDependenceDisplay.from_estimator(
    hgbdt_model,
    X_train,
    **features_info,
    ax=ax,
    **common_params,
    ice_lines_kw={"color": X_train["hour"], "palette": "cool"},
)

The advantage of this is that we get a more detailed picture and can visualize dependencies between the features.
In the first example (all lines in blue) all the plot tells us is that roughly half the lines show a sharp increase in bike rentals (Partial dependence) once the temperature rises above 18°, but it is unclear why some lines show this increase and others not.

With the new functionality, it is possible to dig deeper: The increase in bike rentals almost exclusively happens on the weekend. Possibly because during the week people need to get to and back from work independent of the temperature, whereas on the weekend warmer weather increases the number of bike rentals.

Any other comments?

The example and data is taken from here:
https://scikit-learn.org/stable/auto_examples/inspection/plot_partial_dependence.html

The plots above are also added to this page. The documentation renders well when I tested it, but the output doesn't show up.

I can't understand why the checks regarding matplotlib.colormaps fail, I can run it locally. This was also an attribute that was suggested a DeprecationWarning in matplotlib