What kind of scattergraph, of what kind of data, what's a "dot" in your sentence above, what sort of random distribution, and what "particular area" are you talking about?
If you want to argue science, you're going to have to learn to be far less vague.
Design predicts a scattergraph where the dots are in many very small localized clumps.
Define "very small", "localized", and "clumps" as you are using them in this sentence. Point-like clumps? Linear clumps? Planar clumps? Orthogonal or sloped? This is a multi-dimensional chartspace, is it not? And clumping around *which* kind of locations on the chart? Corners?
Plotting the scattergraph of observed lifeforms on a suitably-dimensional graph
Name the "suitable-dimensions", please.
still yields the second model even today and you know it.
He does? I doubt even *you* "know it". In fact, I think you're just making this up. Feel free to provide a citation to research which has actually plotted any such graph based on real-world data and prove me wrong.
This is a central and essential fallacy of using speciation to teach evolution.
Horse crap. There's no "fallacy" of "using speciation to teach evolution" when it's very heavily supported by multiple lines of evidence, and your imaginary charts can't change that.