As best I can tell, the 5 winning candidates are the ones Clyde outlined. The problem is that we don't actually have an explicitly-established way of counting anymore. In the old game, quota was calculated slightly differently (i.e. the way Clyde referenced above; with an extra "+1" added to the final output), but those rules are no longer in place.
Also, as far as I know, there are no invalid ballots because there are no restrictions on activity requirements unless I've missed a constitutional amendment or something somewhere, so I'm pretty sure there are 48 valid voters.
I just did Paul Lockett on top of that and the numerical outcome is different than Clyde's math (outcome as far as candidates elected is same, though); much messier, but also has Celtic being eliminated before Diptheriadan and has Ben beating out Dip by 0.04 votes. Shows quota at 8.
Another thing to note is that as far as Atlasian precedent goes, I'm pretty sure multiple candidates who hit quota in a given round are elected then and there, rather than doing it one candidate per separate round. Maybe I'm misremembering, though.
Looking at Peebs' updated results - the key difference is that I haven't reduced the weighting for surplus votes. I've done ((candidate transfers/total transfers)*(surplus)) as my calculation for each round - which looking at it now doesn't give a correct result, because of third preferences (and beyond) not having the same transfer value as second preferences, etc.