Tornado chart with colormap gradient

Hi all,

I have created the following code in order to generate a tornado chart with the 'jet' color map gradient with +1 being blue and -1 being red; however, it isn't working! can you please help?

% Data
data = [0.496;  % Electrical conversion efficiency
    0.426;  % Biogas efficiency value
    0.399;  % Biogas potential
    0.386;  % NPK Mineral fertiliser cost
    0.264;  % Methane content
    0.219;  % Volatile solids
    0.009;  % FE
    -0.005;  % Heat Conversion efficiency
    -0.007;  % Capital (Installed) cost
    -0.016;  % Variable O&M
    -0.031;  % GWP
    -0.038;  % POP
    -0.074;  % Fixed O&M
    -0.100;  % AP
    -0.211]; % Parasitic load value
% Sort data
[sortedData, idx] = sort(data);
% Create tornado chart
figure;
barh(sortedData, 'FaceColor', 'flat');
colormap(flipud(jet)); % Blue to Red gradient
yticks(1:length(data));
yticklabels({'FE', 'Heat Conversion efficiency', 'Capital (Installed) cost', ...
    'Variable O&M', 'GWP', 'POP', 'Fixed O&M', 'AP', 'Parasitic load value', ...
    'Methane content', 'Volatile solids', 'Biogas potential', ...
    'Biogas efficiency value', 'Electrical conversion efficiency'});
xlabel('Impact');
title('Tornado Chart');
grid on;

% Color bars based on positive (blue) and negative (red) values
for i = 1:length(data)
    if data(idx(i)) >= 0
        colormapIndex = ceil(data(idx(i))/max(data)*size(jet, 1)/2);
        colorIndex = size(jet, 1)/2 + colormapIndex;
    else
        colormapIndex = ceil(data(idx(i))/min(data)*size(jet, 1)/2);
        colorIndex = size(jet, 1)/2 - colormapIndex + 1;
    end
    colormapIndex = max(min(colormapIndex, size(jet, 1)/2), 1);
    colorIndex = max(min(colorIndex, size(jet, 1)), 1);
    color = jet(colorIndex, :);
    barh(i, sortedData(i), 'FaceColor', color, 'EdgeColor', color);
end

Unrecognized function or variable 'jet'.

Answer 1

[Star Strider]

Perhaps something like this —

% Data
data = [0.496;  % Electrical conversion efficiency
    0.426;  % Biogas efficiency value
    0.399;  % Biogas potential
    0.386;  % NPK Mineral fertiliser cost
    0.264;  % Methane content
    0.219;  % Volatile solids
    0.009;  % FE
    -0.005;  % Heat Conversion efficiency
    -0.007;  % Capital (Installed) cost
    -0.016;  % Variable O&M
    -0.031;  % GWP
    -0.038;  % POP
    -0.074;  % Fixed O&M
    -0.100;  % AP
    -0.211]; % Parasitic load value
% Sort data
[sortedData, idx] = sort(data);
% Create tornado chart
cm = colormap(flipud(jet(numel(data)))); % Blue to Red gradient
figure;
hb = barh(sortedData, 'FaceColor', 'flat');
% colormap(flipud(jet)); % Blue to Red gradient
for k = 1:numel(data)
hb.CData(k,:) = cm(k,:);
end
yticks(1:length(data));
yticklabels({'FE', 'Heat Conversion efficiency', 'Capital (Installed) cost', ...
    'Variable O&M', 'GWP', 'POP', 'Fixed O&M', 'AP', 'Parasitic load value', ...
    'Methane content', 'Volatile solids', 'Biogas potential', ...
    'Biogas efficiency value', 'Electrical conversion efficiency'});
xlabel('Impact');
title('Tornado Chart');
grid on;

% Color bars based on positive (blue) and negative (red) values

% for i = 1:length(data)
%     if data(idx(i)) >= 0
%         colormapIndex = ceil(data(idx(i))/max(data)*size(jet, 1)/2);
%         % colorIndex = size(jet, 1)/2 + colormapIndex;
%     else
%         colormapIndex = ceil(data(idx(i))/min(data)*size(jet, 1)/2);
%         % colorIndex = size(jet, 1)/2 - colormapIndex + 1;
%     end
%     colormapIndex = max(min(colormapIndex, size(jet, 1)/2), 1);
%     % colorIndex = max(min(colorIndex, size(jet, 1)), 1);
%     % color = jet(cm, :);
%     color = colorIndex(i,:)
%     barh(i, sortedData(i), 'FaceColor', color, 'EdgeColor', color);
%     % hb.CData = color;
% 
% end

.

Answer 2

[Voss]

jet is a function that takes up to 1 input argument (the number of colors - call it N) and returns a colormap (which is an N-by-3 matrix).

So you can do this

size(jet,1)

because that calls the jet function with no inputs, in which case jet uses a default value for the number of colors N and returns an N-by-3 matrix, thus size(jet,1) is equivalent to size(jet(),1) which is the number of rows of the matrix returned by jet(), which is N.

But you cannot do this:

jet(colorIndex, :)

because for that to work jet would have to be a variable, but jet is a function.

That was the main problem with the code. To get around it, store the matrix returned by the jet function:

cmap = jet();

and use that where you subsequently use jet.

After changing that, the code ran, but it didn't give the intended result, because a hold on was needed, and also the first barh call can be removed, and the various bars plotted before setting the yticks/labels, etc.

% Data
data = [0.496;  % Electrical conversion efficiency
        0.426;  % Biogas efficiency value
        0.399;  % Biogas potential
        0.386;  % NPK Mineral fertiliser cost
        0.264;  % Methane content
        0.219;  % Volatile solids
        0.009;  % FE
       -0.005;  % Heat Conversion efficiency
       -0.007;  % Capital (Installed) cost
       -0.016;  % Variable O&M
       -0.031;  % GWP
       -0.038;  % POP
       -0.074;  % Fixed O&M
       -0.100;  % AP
       -0.211]; % Parasitic load value
% Sort data
[sortedData, idx] = sort(data);
% Create tornado chart
figure;
% Color bars based on positive (blue) and negative (red) values
cmap = jet();
N = size(cmap, 1);
for i = 1:length(data)
    if data(idx(i)) >= 0
        colormapIndex = ceil(data(idx(i))/max(data)*N/2);
        colorIndex = N/2 + colormapIndex;
    else
        colormapIndex = ceil(data(idx(i))/min(data)*N/2);
        colorIndex = N/2 - colormapIndex + 1;
    end
    colormapIndex = max(min(colormapIndex, N/2), 1);
    colorIndex = max(min(colorIndex, N), 1);
    color = cmap(colorIndex, :);
    barh(i, sortedData(i), 'FaceColor', color, 'EdgeColor', color);
    hold on
end
yticks(1:length(data));
yticklabels({'FE', 'Heat Conversion efficiency', 'Capital (Installed) cost', ...
    'Variable O&M', 'GWP', 'POP', 'Fixed O&M', 'AP', 'Parasitic load value', ...
    'Methane content', 'Volatile solids', 'Biogas potential', ...
    'Biogas efficiency value', 'Electrical conversion efficiency'});
xlabel('Impact');
title('Tornado Chart');
grid on;

It's likely that you can also simplify the logic for mapping each data value to a colormap color, but I didn't change that.