For example it shows EN/DE and EN/NO as having equivalent distances (49) when obviously EN/DE are much closer than either is to NO (in any holistic comparison of the three languages). It also shows IT/FR as being significantly closer (31) than IT/SP (40) when that's also plainly not the case -- again if we consider the distance between respective language pairs holistically (which includes a whole lot of other factors like phonemic footprint, meter, etc).

In short this reads like pretty much the kind of post one would expect from a "brain candy" site like this.

Less harshly (assuming good indent on the author's part), it demonstrates the pitfalls of relying on a single score (which in itself can be subject to all kinds of weighting and sampling biases) simply because it's a "score" and we're all supposed to be "data-driven" in our analyses these days.

But again -- nifty diagrams, and they do at least give a feel for the topic of mutual intelligibility and language distance (even if the scores themselves don't seem to be particularly meaningful).