It seems to me that it's affine transformation, not linear, insofar as we consider the drawn intersection point of the axes to be the zero point of display-space. (It would be linear if the data were on the left of the axis.)

(Context: I am currently taking a course on linear algebra, so this came readily to mind, but I may lack further relevant information.)

Also, looking at the slides, the x axis is increasing rightward in its (unfortunately black-on-dark-blue) labels. So it's not so much the graph as the scale being used that is flipped. Perhaps they simply plotted the data using default sort-numerically-increasing software settings and didn't think about it too hard since they're used to working with that scale.