Is there an R package that easily identifies an observation that best "represents" the model?"

Great. Please mark the solution for the benefit of those to follow. Do this even for your own solution, no false modesty. And, so far as I know,

a penny a point, ain't no one keeping score