Sorry to go on at this subject but coming from an Engineering background I find demography interesting. Also as an educator I am always having to impress on Students that finely drawn graphs etc are no better than the underlying data upon which they are based - knowledge of sources and their limitations is I'm afraid not just a sin of University students but also of many family historians.

The webpage gives a claimed distribution for some 6200 family names, no explicit source is given. The 2002 census has just under 32,000 households thus 6200 would appear to be a 20% sampling of the target population.
Taking a couple of names as checks - Leece is stated to occur 37 times thus one would expect about 200 Leece households in the Island. Most Manx families have a telephone and due to the absence of annoying telephone sales companies they generally keep their number in the directory - in 2002 there are some 46 non-commercial numbers once duplicate mobile numbers removed - telephone penetration is generally thought around 90% thus there is an immediate discrepancy - estimated households from Directory 50-60; estimated from the webpage 200. The discrepancy is futher compounded by looking at distribution of those households in the Telephone directory - 20% in Peel, 10%(approx) in German, Braddan, Onchan, Douglas, Ramsey and Rushen ; 5% in Arbory, Lonan and C'town - quite at odds with the published map.

Likewise in Directory there are 5 columns of Kellys and just over 2 of Smiths - again major discrepancy
Taking the distribution of Kelly households - generally Kelly is thought of as ubiquitous - I took the 1881 census, extracted all males aged 50 or older and looked at their claimed birth parish (ie a distribution dating from 1831) - all towns and parishes were represented 12% in Douglas, 10% in German, Marown and Patrick; 4-8% elsewhere except Arbory, Andreas, Bride, Juby,Lezayre, Ramsey and Santan contributed less than 3% each. Thus relying on the published map as to where to search for records is not a sensible guide (for example applying this to Leece name yields Arbory as supplying 50% but this would be incorrect German, Patrick and Santan also have many.
As a possible locator of unusual names the published maps may have a use but as a locator guide to records ?

Finally the Moral of that, as the Red Queen would say, please be suspicious of sources especially those that give no provenance to check on - they may be offered with the best will in the world but that does not unfortuneately carry any garantee.