You can find the source that generated the graph here: https://github.com/rinze/hn_remote/blob/main/hn_remote_main....

I didn't go through the raw dataset (which is generated by this query, https://news.ycombinator.com/submitted?id=whoishiring) but two obvious issues with the count are, it didn't filter out duplicates and it's just grep'ing the post for the word remote. So for example, "remote work not allowed," counts as a posting for remote work.

I have no doubt the demand for remote work is up but I am not sure this count accurately measures by how much, even for just hacker news postings.