Besides, the direct short-term deaths from commercial nuclear nearly all came from Chernobyl first responders, where staff and firefighters were the ones who got ARS (they chose hazardous careers too).
For the "up to 4000" latent deaths from Chernobyl, I think you have a good point. 1960s nuclear is worse than wind and solar in latent deaths. I'm hoping advanced nuclear with passive decay heat removal and low pressure coolants will even that out.
I also didn't mention the reasons behind deaths from wind, which include hazardous maintenance and ice throw. Deaths from these things are exceedingly low per kWh, just like they are for nuclear. That's the point I'm trying to convey: nuclear is roughly on par with wind/solar in terms of deaths/kWh.
Anyway, where is the best data on what kind of solar is currently being built?
One could make the case that the public is also subject to risk from solar, for example from traffic accidents transporting the materials to and from large scale installation sites. But the "falls from roof tops" argument isn't doing that.
People often meet this idea wish shock: "But how can wind or solar hurt anyone? There are no Captain Planet episodes about that." So I have to explain some mechanisms of why huge distributed energy harvesting can harm people. If you have other suggestions on how to communicate this effectively I'd love the help.
First, most PV installs in the US are now large scale, where the PV modules are installed on the ground, not on roofs.
Second, deaths of workers in an industry are different than deaths of the general public. Workers made a choice to work on roofs; the public has no choice if a reactor near them has a problem.