You need to scale each parameter (MPG, Horsepower, etc.) by the maximum value of each parameter. I used repmat in my code so that the scaling could be done in a single line instead of a for-loop.
% Find maximum value for each variable
% Use repmat to allow for vectorization (faster computation)
column_max = repmat(max(data), size(data, 1), 1);
% Scale data so that max value for each variable = 1
data_scaled = data./column_max;
h = figure();
set(h,'DefaultAxesColorOrder',[1 0 0; 0 1 0; 1 1 0; 1 0 1; 0 0 1]);
lineobjects = parallelcoords(data_scaled,'Group',Cylinders,'Labels',labels);
The code above will produce the following figure (not an exact replica of the figure you sent, but you can mess around with annotations and axis labels):
Hope this helps!
