The killings that nuclear may make are difficult to pin down (did that cancer come from nuclear exposure, or a chemical, or just a chance cosmic ray?) and so may be unattributed.
A well-developed statistical model should be able to take that into account. My understanding is that the academic consensus is fairly strongly leaning towards nuclear being the safest energy option, based on robust statistical models. Would be interested in hearing the opinions on this from someone actually in this academic field though...
A model could provide an estimate there, but nuclear deaths are not estimated that way from the posted source. They are only directly attributable deaths. e.g. cherynobyl was attributed from the posted source as 7 deaths while statistical modeling puts it in at least the tens of thousands range. Even the 7 of the link is questionable as reports of firefighter deaths during the accident as 27.
Skin cancer can also come from exposure to material on the skin or ingested... chemical or nuclear, it could come from strong enough exposure to EM radiation. at the point you have cancer, it's very difficult to say exactly what the cause was.
