You'll want to have a stronger link before you can worry about what kind of relation it is exactly. And if you ignore the people with scores <60% you lose most

You'll want to have a stronger link before you can worry about what kind of relation it is exactly. And if you ignore the people with scores <60% you lose most of the correlation entirely, which to me suggest there's something causing both bad sleep and bad academic performance rather than there being some continuous relation. In fact the distribution looks bimodal which suggests there's something splitting the population in 2.