I think it would be nice to add the formula for binary balanced accuracy. It's in the text right now, but maybe adding the formula would be nice as well.
The text says

In the binary case, balanced accuracy is equal to the arithmetic mean of sensitivity (true positive rate) and specificity (true negative rate), or the area under the ROC curve with binary predictions rather than scores.

It should be something like

$$\text{balanced_accuracy} = \frac{1}{2}\left( \frac{TP}{TP + FN} + \frac{TN}{TN + FP}\right )$$

I think.