``mapping.nonlinear_fit``

[Sibgatulin]

The function can be used to illustrate how an attempt to write concise code on somewhat higher level of abstraction can come at such a high performance price that the whole attempt becomes completely impractical.
I have a (strong) suspicion that the function should be torn apart and stepped through in Jupyter otherwise the example is useless for those unfamiliar with pandas.

[Sibgatulin]

I suppose, if this example is shown at all, it would benefit from slides walking through key ideas (groupby-apply and the fit function). Yet out time constrains will likely push such a specific example out of the 10 mins time slot