Reduction of the impact of the local valley splitting on the coherence of conveyor-belt spin shuttling in 28Si/SiGe

v2: Very slight correction of one valley splitting jump that was read off wrongly in v1 of the dataset at (y,d)=(-12nm,133nm). It was not clear at which precise nanometer the upper line ends and the lower line starts. When zooming in and reading off, one point at the top was put a bit too far too the right and one at the bottom too far to the left, resulting in a retrograding spline which is unphysical and simple to correct. It is corrected now and the data is algorithmically checked for other points like this. v3: Correction of the two dimensional valley splitting map creation from the one dimensional traces.