I was looking for a plot function in Matlab that could create error “bands” with a continuous filled plot for the standard error, highlighting the mean. In the end I decided to create the function I needed so that I could have the parameters I wanted: namely colours and transparency.
Here an example of how to use two versions of the same function, one for matrices and the second for a structure of data. Download the archive with the two functions.
|
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 |
clear all close all figure('Position', [200 200 1500 500]); %% test with consistent sample of data data_mat1=rand(10,20); %first set of data: mean and standard error are calculated per column data_mat2=rand(10,20); %second set of data: mean and standard error are calculated per column color1=[0.1 0.7 0.3]; %color used for the first set of data color2=[0.2 0.2 0.7]; %color used for the second set of data t1=1.0; %transparency of the first fill plot (1= solid, 0=transparend) t2=0.5; %transparency of the second fill plot (1= solid, 0=transparend) subplot(1,2,1) filled_plot(data_mat1, color1, t1, data_mat2, color2, t2) %% test with uneven sample of data for i=1:20 data_mat3{i}=rand(10+i*10,1); %first set of data: mean and standard error are calculated per cell data_mat4{i}=rand(10+i*10,1); %second set of data: mean and standard error are calculated per cell end color3=[0.5 0.5 0.5]; %color used for the first set of data color4=[0.5 0.1 0.1]; %color used for the first set of data t3=0.8; %transparency of the first fill plot (1= solid, 0=transparend) t4=0.4; %transparency of the first fill plot (1= solid, 0=transparend) subplot(1,2,2) filled_plot_cell(data_mat3, color3, t3, data_mat4, color4, t4) |
Function 1 (matrix of values)
|
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 |
function fig= filled_plot(input_vecs1, color_mean1, transparency1, input_vecs2, color_mean2, transparency2) mean_vec1=mean(input_vecs1); mean_vec2=mean(input_vecs2); error_vec1=std(input_vecs1)/sqrt(length(input_vecs1(:,1))); error_vec2=std(input_vecs2)/sqrt(length(input_vecs2(:,1))); color_error1=min(1, color_mean1+0.3); color_error2=min(1, color_mean2+0.3); x1=1:length(mean_vec1); x2=1:length(mean_vec2); X1=[x1,fliplr(x1)]; X2=[x2,fliplr(x2)]; y1=mean_vec1+error_vec1/2; y3=mean_vec2+error_vec2/2; y2=mean_vec1-error_vec1/2; y4=mean_vec2-error_vec2/2; Y1=[y1,fliplr(y2)]; Y2=[y3,fliplr(y4)]; hold on fig{1}=fill(X1,Y1, color_error1, 'EdgeColor', color_error1); fig{2}=plot(mean_vec1, 'Color',color_mean1); fig{3}=fill(X2,Y2, color_error2, 'EdgeColor', color_error2); fig{4}=plot(mean_vec2, 'Color',color_mean2); set(fig{3},'facealpha',transparency1); set(fig{3},'facealpha',transparency2); hold off end |
Function 2 (structure of values)
|
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 |
function fig= filled_plot_cell(input_struct1, color_mean1, transparency1, input_struct2, color_mean2, transparency2 ) for i=1:length(input_struct1) mean_vec1(i)=mean(input_struct1{i}); mean_vec2(i)=mean(input_struct2{i}); error_vec1=std(input_struct1{i})/sqrt(length(input_struct1{i})); error_vec2=std(input_struct2{i})/sqrt(length(input_struct2{i})); end color_error1=min(1, color_mean1+0.3); color_error2=min(1, color_mean2+0.3); x1=1:length(mean_vec1); x2=1:length(mean_vec2); X1=[x1,fliplr(x1)]; X2=[x2,fliplr(x2)]; y1=mean_vec1+error_vec1/2; y3=mean_vec2+error_vec2/2; y2=mean_vec1-error_vec1/2; y4=mean_vec2-error_vec2/2; Y1=[y1,fliplr(y2)]; Y2=[y3,fliplr(y4)]; hold on fig{1}=fill(X1,Y1, color_error1, 'EdgeColor', color_error1); fig{2}=plot(mean_vec1, 'Color',color_mean1); fig{3}=fill(X2,Y2, color_error2, 'EdgeColor', color_error2); fig{4}=plot(mean_vec2, 'Color',color_mean2); set(fig{3},'facealpha',transparency1); set(fig{3},'facealpha',transparency2); hold off end |