This has been suggested before here: https://www.reddit.com/r/toolbox/comments/4fnyvp/sentiment_analysis_for_highlighting_aggressive/

This site is doing something similar with a kindness score based on reddit controversiality and bad vs. good words: https://atomiks.github.io/reddit-user-analyser/

The code driving the score: https://github.com/atomiks/reddit-user-analyser/blob/master/src/components/UserSummary.vue#L242

You can see it is perhaps too simple with its word list, but it does seem to work. I ran it against a set of banned users, etc. and it repeatedly reported 0%. A couple things that could be improved would be to expand the word lists and perhaps have a "unknown" for users with too short of a history.

It seems like this would fit in the user history module since it is already pulling back user comments and submissions. In fact I don't really use the history module because I'm not overly concerned with spammers (which the history module seems to focus on) but I would use something like this regularly.