Getting complications (allergy, ...) from the flu shot should be included somewhere in the graph

The Coleman paper provides some stats on how likely e.g. muscle pain is to result from receiving a flu shot (5-30%, authors are not sure how significant this effect is, though a couple RCTs report this effect), as well as on fever (which 3.2% of those who got a flu shot, and 1.6% of those who received a placebo experienced, according to one RCT cited by Coleman).

I never saw anything about the probability of experiencing an allergic reaction to the vaccine, but I haven't specifically looked for this information.

But most importantly (and this time, arguing for the shot), the "herd immunity" effects should be considered.

Yes, they could be. The reason I didn't consider herd immunity effects is because I was calculating the benefit which a (possibly un-altruistic) individual would receive from vaccination.