Yes, you can call scatter and use the fourth input, which is colors.
It is not necessary to use anonymous functions for this, because bitand and bitshift accept array inputs.
% a random table of data:
rfs_csv_table = table((1:10).',randi(65535,10,1,'uint16'));
% some fixed i, for demonstration:
i = 2;
% temporary variable containing contents of ith column:
tmp = rfs_csv_table{:,i};
% method 1: the anonymous function route:
fktPRN = @(sigID) bitand(uint16(sigID),uint16(255));
fktSatType = @(sigID) bitshift(bitand(uint16(sigID), uint16(7936)),-8);
tmpPRN_anon = arrayfun(fktPRN,tmp);
tmpSatType_anon = arrayfun(fktSatType,tmp);
% method 2: calling bitand and bitshift with vector input,
% instead of using anonymous functions:
tmpPRN = bitand(uint16(tmp),uint16(255));
tmpSatType = bitshift(bitand(uint16(tmp), uint16(7936)),-8);
% the results from the 2 methods are the same:
isequal(tmpPRN_anon,tmpPRN)
isequal(tmpSatType_anon,tmpSatType)
% let's see the values, for verifying that the scatter plot looks right:
disp(tmpPRN)
disp(tmpSatType)
% make a scatter plot, with tmpPRN as y and tmpSatType as colors:
scatter(rfs_csv_table{:,1},tmpPRN,[],tmpSatType,'filled');
colorbar
Also, note that if the columns of data in the table are of type uint16, then there is no need to cast them to uint16 using uint16(tmp) (or uint16(sigID) in the anonymous functions), but I kept it that way here because that's how you had it.
Also, if the width of the table doesn't change inside the loop, then there is no need to check if((i) <= width(rfs_csv_table)) because i will always be less than or equal to the width of the table since that's how the for loop is defined: for i = 36:(width(rfs_csv_table))
