Dashboard plots (fairness) - Use user performance metric

[michaelamoako]

The Fairness dashboard currently shows selection rate and fpr/fnr for binary classification. It would be useful if instead of these, the displayed plot depended on the user's chosen metrics (i.e: if Recall is the chosen metric, the plot below should be for Recall)

[michaelamoako]

The above is achievable without the dashboard using the following code (similar to @romanlutz, fairlearn/fairlearn#766):

A. metric_frame.by_group.plot(kind='pie', y='Count', figsize=[12,8])
B. metric_frame.by_group.plot(kind='bar', y='Recall', figsize=[12,8])

Where metric_frame is a MetricFrame that includes Recall and Count (defined by the user). I separate the plots because for Recall for example I would like a bar chart, while for Count I would like a Pie chart. If the same customizability could be added to the dashboard (instead of default selection rate, fpr/fnr) that would be very useful.

(Documentation here: https://pandas.pydata.org/pandas-docs/stable/reference/api/pandas.DataFrame.plot.html)