Even if you assume that all Muslims have positive views about Muslims, the correlation in the graph still stands. Eg France would go from 22% anti-Muslim to 24% (and the countries lower down would barely change).

European ones were, given this was a European survey. (Though given Europe managed to murder most of its Jewish population I expect there weren't many Jews among the respondents.)

First comment was semi in jest. Though these sort of numbers are culturally dependent and not that easy to compare: what people say and how they act don't always align.

The second comment is true (in fact the 2016 numbers were recently posted on mapporn which is what gave me the idea to do this).

Fair point. I'm still experimenting with this type of graphing. I might try using circular flags (more accurate as points) with either the full country name or an abbreviation.

My theory is that Greeks mainly hate pollsters.

That said, the numbers are both down over a similar poll in 2016 (three years before).

Sources: https://www.pewresearch.org/global/2019/10/14/minority-groups/, https://en.wikipedia.org/wiki/Islam_by_country and https://en.wikipedia.org/wiki/Jewish_population_by_country. Made using matplotlib.

The Pew report also included data about attitudes to Roma, but I couldn't find reliable population figures.