Fair! Like you say, there's still gonna be a score "underneath" that dictates whether the label should be shown or not, and defining any such threshold is always more art than science too, but I get your point.

My hope is that we won't reinvent the wheel all the time. A lot of good research exists in the world of graphs and networks, and these centrality measures are solid. We'll make a lot of mistakes and miss lots of good stuff if we try to come up with our own naive metrics for everything.