First, very happy to see someone using data of this kind to answer this question! While I am going to critique things a bit here, please do not take this as any kind of discouragement or putting-down. 
I see you say you’re not a data scientist or psychologist, so I’m not expecting super-rigorous analysis of what appears to be something of a passing thought here.
Second, while your general drawing on theory etc is certainly a good thing, I think your connection to the work around virality is slightly inappropriate. Virality does tend to both elicit and result from strong emotions including anxiety and anger, but it’s not the same as those things (although it’s definitely plausible it’s correlated, so I guess you could use it as a kind of proxy if you needed to). I see things that make me anxious or angry all the time, but I don’t necessarily share them and I very much doubt any have gone viral even among my friend group if I do share them. A better way would be to have noted at the time how a post made you subjectively feel - I’ve read a lot of articles that were clearly trying to make me sad or anxious or angry and I’ve just found them hilarious, for example! But if you read a (ideally random, but that’s not strictly necessary) selection of posts with various tags, then rate how you feel on various emotions (maybe on a spreadsheet?), then you’d be better placed to make that judgement. As it stands all you can say is:
- You engage with a lot of bad-news tagged posts (a disproportionate amount? You’d need to look at how many posts are tagged of all posts, but you could probably eyeball this if it’s a really strong effect), as measured by your use of various reactions.
- I note that you used the thumbs-up reaction as often as you used the stressed face reaction. One of those is positive, while the other, yes, likely indicates anxiety of some sort. And since neither are the default reaction (that’s the heart, I believe?) that indicates an active choice to use it, so I wouldn’t rush to the conclusion of too much anxiety. Your second-most used reactions were the celebration reaction and the rofl reaction - both very positively-coded, I would say! Although apparently there’s a reasonably strong sex difference here, which I didn’t know. Kind of cool work, I’ll have to dig deeper into it.
The article you cite also is doing something a little differently than you’re talking about. They took articles and measured the virality (in a way which I have serious doubts about, but that’s a whole other discussion) as related to the frequency of various emotionally-coded words within them. As noted above, we’ve all read posts/articles which use a lot of emotional language, only for our reaction to be very different, so this isn’t exactly a great proxy for your purpose.
In addition, they note that articles that were more positive were, all else being equal, more likely to appear on the ‘most e-mailed’ list than more negative ones. As they note:
The results indicate that content is more likely to become viral the more positive it is … This suggests that transmission is about more than simply sharing positive things and avoiding sharing negative ones. Consistent with our theorizing, content that evokes high-arousal emotions (i.e., awe, anger, and anxiety), regardless of their valence, is more viral.
So more emotionally laden terms means more viral overall, but more positive words is better than more negative words, all else being equal.
Your point about negativity bias is a good one, It’s generally well-established, but there is some interesting work which suggests it’s more common in younger people than older ones. Which anecdotally sounds right - the younger people I know are way more prone to doom-saying and worrying than the older people.
Anyway, I’ll stop now. Those were just my immediate thoughts, but like I said I’m really happy to see you using research and data and stuff to try to understand this! Keep it up!