Interesting. You are asking the same question as Ms. Wen Xue asked: https://www.researchgate.net/post/Is_this_practice_a_p-hacking_or_the_garden_of_forking_path

Thanks for your reply. Because this question is very important and interesting. I hope more people will participate in this discussion and give their opinions. Hening Huang

It certainly would not be sound to publish this result as if you had had that "hypothesis" prior to looking at the data. It may be fine to publish it as an explicitly exploratory (data-generated/data-informed) finding with a cautionary note stating that this is a data-driven result that needs to be validated/replicated with fresh (and ideally experimental or at least quasi-experimental) data (truly experimental may be difficult in this case because it may not be possible or ethical to manipulate life satisfaction).