The 'dataset' file is split in two. 

First, 'Experimental' contains the data for the AFM plot in figure 1 and the far-field data for figure 2. 
AFM image: skip the first row of the 'U-shape.txt' document (headers) and use the first column 'height' to plot the AFM image with the software of choice. 
In order to plot three-dimensional image, create  2D matrix of height with no. columns and rows = sqrt(len(height)) and plot against x and y coordinates (create array from 0 to 4 with 'sqrt(len(height))'-step size).

Far-field: 
For each nano-array there is a separate folder. Each folder contains 330 data files (csv format). The number-suffix represents the diffraction angle for which the spectra was taken. The suffix 'M' states that LCP illumination was used for this spectrum, while 'P' is RCP.
To plot the CD-Diffraction patterns shown in figure 2, do CD = (RCP-LCP)/(RCP+LCP) to calculate the CD at every point of the spectrum for each angle. 
Plot result vs 'diffraction angle' (from filename) and 'wavelength' (from file).

The 'Numerical' folder contains all data used for the numerical part in figures 2 and 3. 
The 'Far-field' folder contains data for every array, as for the experimental part above. However, the data is saved in '.dat' file format. The suffix 'neg90' is for LCP illumination and 'pos90' RCP. To plot CD data as scatter plot, repeat maths from before. 

The 'Near-field' folder is split into the data for the maximum and minimum CD response (in far-field), as 'Max CD' and 'Min CD' respectively. Each contains six '.dat' files, 3 each for LCP and RCP illumination, measuring the individual components of the electric field at the surface of the arrays. 
x and y are both varying from -1.2 um to 1.2 um. The real part of the electrical field stores in the (2k-1) columns, while the imaginary part in the (2k) columns (k=1,2,...,513). To plot the electric near-field, sum all intensities, i.e. all the squares of the complex E-fields and plot vs. x and y.