Hey i just saw your video on solving, it was pretty good but in some cases i would do something different, (that is because i main the Roux method).
that aside for the beginners method you did do quite good but kinda sloppy, i noticed that after you inserted 1 cross piece you stopped to look for the other cross edges, this could make your solve more slower, you should use your 15 second inspection time on planing the entire white cross or at least planning the "Daisy" that you make before solving the cross edges.
For the bottom layer corners and the middle layer edges, you did well. but you rotate the whole cube at where the corners and/or edges are, this could be because you were seeing the cross edge. I would recommend putting your cross at the bottom side of the cube since after the cross is done you don't really need to think or look at the cross and by doing this you could improve your "look ahead".
now if you don't know what is "look ahead". Look ahead is for example when you put your bottom layer pieces and you instantly continue to solve the other pieces, how do you do this? an example on how to do this is when you insert your corner you shouldn't really think too much about inserting it, rather while you insert it you should also look at the top layer to see if you could spot a corner for your next insertion.
now for the last layer i would recommend learning 4 Look Last Layer, 4LLL is when you solve the top layer by doing 4 steps. If i remember correctly 4LLL have 12 algorithms.
other thing that you should improve are stopping to look for the pieces, while everybody does that, it would cause your solve to be more slower than you could potentially be, to counter this you would need a fairly good understanding on how to do "look ahead". i personally aim for clean solves, and what i mean by Clean is i am aiming to solve the cube with no pausing, rather than i solve fast but i pause like 100 times to find the corners or edges to solve.
Now Please Don't Think Of My Tips Very Seriously Since I Am Also A Beginner Roux solver, And I Average Of 43.27 Seconds Per Solve.