You need to compare the raw files on the DPReview comparator, and if you do, you will see that there isn't much difference visually with noise in the deepest shadow areas (which is where it counts).
Personally I don't pay any attention to the DXO Mark scores because how they arrive at them is opaque.
Photons to Photons is the one to pay attention to but the numbers in the charts you link to are derived from the DxO data.
Here is another chart showing readout noise at various ISOs comparing the two cameras from Photons to Photons where I think the data is original. This is the one I would pay the most attention to.
It shows the Canon with 3.555 e- of read noise at ISO 200, and the Nikon with 2.732 e-.
That is a little bit of a difference in the Nikon's favor.
Unless you are shooting under exceptionally dark skies with either slow focal ratios or filters, this probably won't make much of a difference.
Between these two cameras, I would pick the one you like the most in terms of the features and ergonomics.
The quality of the long-exposure deepsky images produced with them will almost certainly come down to the photographer's skill and expertise than the camera's technical capabilities.
That said, if you are really serious and plan to shoot calibration frames, the D5300 is kind of weird with flats with some lenses. And some of the recent Canons have had problems with artifacts from the autofocus pixels. You should research this here on Cloudy Nights where it is documented, if not actually solved.
"One road leads to despair and utter hopelessness and the other to total extinction." Be sure to choose wisely. Ok, so it's not quite that bad... :-)
Edited by Jerry Lodriguss, 14 June 2018 - 12:29 PM.