I think we have to be careful with all the "Sizzi" samples. They're by no means representative, having been chosen by him from who knows where and for who knows what purpose.
That said, I don't think Calabria has some of the "northern" input which went into some areas of Sicily. The additional Saracen/North African input which the Sicilians have was counterbalanced by that Northern Italian input at least as far as placement on a PCA goes.
I touched on that looking at my own results. I've never gotten a "match" with western Sicilians, i.e. at 23andme or Ancestry. However, I've seen other signs of some sort of connection between Sicily, at least the more central and northwestern parts of Sicily, and northwestern Italy. I think it relates back not only to the Lombard settlement, when many Ligurians, Tuscans, Lombards, were settled there, but also to ancient settlement patterns in Sicily. None of that applies to Calabria.
Northeastern Sicily is a bit different. I think they would probably plot closer to Calabrians, as they were less influenced both by the Saracens and the "Lombards", and parts of it, Messina, for example, were settled from Calabria after the major earthquakes there.
As for Campania, I think there is variety there because of the presence of Napoli, where there has been more admixture, and the hinterland, famously near where the expelled Ligures were settled.
Napoli, like northwestern Sicily, may have more "northern" ancestry, including French ancestry, but some expelled Saracens were also settled there, near Salerno, "Saracens" who may have been at least partly native Sicilian.
One just has to look at the variation in Toscana on the PCA above, which had a far less turbulent history, to see how much variation there can be in a single province in Italy. One can see it in Bergamo as well. What is interesting is that if you and Jovialis are representative of Puglia, it is more homogeneous, unless, of course, one of you had your kit used to create the Puglia sample.
Of course, this is just conjecture, although, I hope, based on data and history. I would need lots of samples and the ability to compare them to say anything more certain.
It would be interesting, for example, to plot the samples from this paper on the PCA. They are from northwestern Italy, i.e. the Ligurian hinterland which is now part of Piemonte, an area in the isolated northeast, and Foggia in Puglia.
Maybe Jovialis knows if the samples are available.
This is the paper;
https://www.nature.com/articles/s41431-019-0551-x
It would also be interesting to see the samples from the Corsica paper on the plot and see how homogeneous, or not, they are...