Was 2018 the Year of the Graph Database?

Cost, I think.

I don't disagree at all, but query languages like "SPARQL" and "Cypher" are no where nearly as approachable as the query by example like syntax of "GraphQL" (resolvers aside).

My feeling is that GraphQL has opened up a pathway and other than that I personally did not observe any other kind of sea change in the graph database landscape.

Because the article is not much more than an elaborate wall of buzzword bingo and name-dropping. Nowhere does it answer the question it imposes on itself, and explains how a "yes" could be a serious answer to the question. Where they mention 2018, they talk about the knowledge graph, not graph databases. Nowhere does the text go into depth about why traditional database might not be suitable to the task of "[combining] quantitative and statistical bottom-up analysis with qualitative analysis based on expert-provided knowledge". It's a text-book example for Betteridge's law.

I think the criticism is quite clear. Asking a question as a headline is a form of clickbait. If the statement implied by the question was a verifiable fact, it would be posed as such instead of lying by phrasing it as a question.

It's already a critique of the cliche being used in the headline. Use of this rhetorical device already casts doubt upon the veracity of the content therein, not to mention the skill of the writer.

Also, ironically enough, by posing a question in the headline, the answer of said question is automatically "no." You've now saved yourself the time you'd otherwise have wasted reading the article.

Although you may think that the comment pointing out Betteridge is a tired meme, you should really be ridiculing the writer of the article for utilizing the tired meme of putting unverifiable information into the world by writing a headline as a question.