More Realistic Data
Real data is not just numeric and if you have a number of subjects, or sampling points, there is a good possibility that you will need more than one observation from each. The most general way to handle all possibilities is to put the results in a structure. Here are some structure data you can cut and paste:
position(1)=struct('name', 'Digford', 'grid', [010 020], 'temp', [68, 80, 72], 'rainfall', [0.2, 0.4, 0.0]);
position(2)=struct('name', 'Lagton', 'grid', [013 015], 'temp', [64, 83, 70], 'rainfall', [0.4, 0.2, 0.2]);
position(3)=struct('name', 'Beerness', 'grid', [017 022], 'temp', [81, 42, 32], 'rainfall', [0.0, 0.0, 0.2]);
and here is a script (or function if you want to write it as a function) that uses this data:
for ii = 1:length(position)
subplot(2,3,ii);
bar(position(ii).temp)
title(position(ii).name)
ylim([0 100]);
subplot(2,3,ii+3);
bar(position(ii).rainfall,'r')
ylim([0 0.5]);
end
Again, the point here is for you to understand how this works and to make changes to see if you can make it work better. Like this:
data = zeros(3,2);
for jj = 1:3
for ii = 1:length(position)
data(ii,1) = position(ii).temp(jj);
data(ii,2) = position(ii).rainfall(jj)*100;
end
subplot(1,3,jj);
bar(data)
legend({'Temp', 'Rainfall * 100'}, 'Location', 'SouthOutside');
title(['Day ' num2str(jj)]);
end