Skip to content
A repository comprising multiple functions for making elegant publication-quality figures in MATLAB
MATLAB
Branch: master
Clone or download
Latest commit e73c72b May 29, 2019
Permalink
Type Name Latest commit message Commit time
Failed to load latest commit information.
Boxplot Update figure_boxplot.m May 29, 2019
GeneratePDF Update generatePDF_examples.m May 3, 2018
Heatmap Revised figures May 9, 2018
README.md Update README.md May 9, 2018

README.md

Making Elegant Matlab Figures (in progress)

A repository comprising multiple functions for making elegant publication-quality figures in MATLAB

Table of Contents

Boxplot

Generating boxplot in MATLAB using the default function boxplot.m is a bit cumbersome due to the large number of required (and somewhat strict in terms of format) inputs. Here, I have written a wrapper code for making nice boxplots quickly and efficiently.

Function: figure_boxplot.m

Example 1: Boxplot using only data input and no other specification

% Number of intended boxes in the figure
num_boxes = 8;          

% Generating random data
data = cell(1,num_boxes);   
for k = 1:num_boxes
    data{k} = randi(10) + randn(1,1000);
end

% Using the "figure_boxplot.m" function to plot the boxplot figure using the data

figure_boxplot(data)

alt text

Example 2: Boxplot using minimum input specifications

% Number of intended boxes in the figure
num_boxes = 6;          

% Generating random data
data = cell(1,num_boxes);   
for k = 1:num_boxes
    data{k} = randi(10) + randn(1,1000);
end

% Using the "figure_boxplot.m" function to plot the boxplot figure using the data, 
% x- and y-axis labels, and label of each box.
% For more information related to function inputs, check the function "figure_boxplot.m"

label_axes = {'Variable','Number'}; 
label_boxes = {'alpha','beta','gamma','delta','epsilon','zeta'};
figure_boxplot(data,label_axes,label_boxes);

alt text

Example 3: Boxplot using data input in matrix format

Instead of a cell, the code works even if the input is in matrix form of size num_samples x num_boxes.

% Number of intended boxes in the figure
num_boxes = 7;  
% Number of samples in each box plot
num_samples = 1000;

% Generating random data
data = zeros(num_samples,num_boxes);   
for k = 1:num_boxes
    data(:,k) = randi(10) + randn(num_samples,1);
end

% Using the "figure_boxplot.m" function to plot the boxplot figure using the data

figure_boxplot(data)

alt text


GeneratePDF

This is a code to generate nice (properly normalized) probability density function (PDF) plots with minimum amount of input arguments.

Function: generatePDF.m

Example 1: Generating PDF with only data input

% Generating random data
x = randn(1,500); 

figure;
generatePDF(x)

alt text

Example 2: Generating PDF in a specified color

% Loading colors
run colors_definitions.m

% Generating random data
x = randn(1,500); 

figure;
subplot(2,1,1)
generatePDF(x,'b') %You can use any color
title('Using one of the default colors')
subplot(2,1,2)
generatePDF(x,color_scheme_set1(1,:)) %You can use any color
title('Specifying a color manually')

alt text

Example 3: Generating PDFs in different styles

% Loading colors
run colors_definitions.m

% Generating random data
x = randn(1,5000); 

figure;
subplot(3,1,1)
generatePDF(x,color_scheme_set1(1,:),'hist') 
title('Histogram plot')
subplot(3,1,2)
generatePDF(x,color_scheme_set1(1,:),'curve') 
title('Curve plot')
subplot(3,1,3)
generatePDF(x,color_scheme_set1(1,:),'area') 
title('Area plot')

alt text

Example 4: Generating PDFs with specific number of bins

% Loading colors
run colors_definitions.m

% Generating random data
x = randn(1,5000); 

% Specifying the number of bins
no_of_bins = [20 30 50];

figure;
for k = 1:3
    subplot(3,1,k)
    generatePDF(x,color_scheme_set1(2,:),'hist',no_of_bins(k))
    title(sprintf('Bins = %d',no_of_bins(k)))
end

alt text

Example 5: Generating overlapping PDFs

% Loading colors
run colors_definitions.m

% Generating random data
x = randn(1,1e4); 
y = 2 + randn(1,1e4); 

figure;
generatePDF(x,color_scheme_set1(1,:),'area')
hold on
generatePDF(y,color_scheme_set1(2,:),'area')
legend('Data 1','Data 2'); legend boxoff

alt text

Example 6: Generating PDFs and saving figures

% Loading colors
run colors_definitions.m

% Generating random data
x = randn(1,1e4); 

% Specifying the number of bins
no_of_bins = 50;

% Data for saving figure in png format (4 inputs required)
savefig = 1; % 1 --> you want to save figure 
fig_name = 'generatePDF6';
fig_width_cm = 16; 
fig_height_cm = 10;

figure;
generatePDF(x,color_scheme_set1(3,:),'area',no_of_bins,...
    savefig,fig_name,fig_width_cm,fig_height_cm) 

alt text

Example 7: Generating overlapping PDFs with different color schemes

% Loading colors
run colors_definitions.m

% Generating random data
num_samples = 2e4;

x(1,:) = randn(1,num_samples); 
x(2,:) = 2 + 1.25 * randn(1,num_samples); 
x(3,:) = 4 + randn(1,num_samples); 
x(4,:) = 6 + 0.9 * randn(1,num_samples); 
x(5,:) = 8 + 1.5 * randn(1,num_samples); 
x(6,:) = 10 + 0.9 * randn(1,num_samples); 
x(7,:) = 13 + 1.1 * randn(1,num_samples); 
x(8,:) = 16 + 0.9 * randn(1,num_samples); 

% Color schemes to test
no_color_schemes = 5;
color_scheme{1} = color_scheme_npg;
color_scheme{2} = color_scheme_aaas;
color_scheme{3} = color_scheme_nejm;
color_scheme{4} = color_scheme_lancet;
color_scheme{5} = color_scheme_set1;

titles_schemes = {'NPG color scheme','AAAS color scheme',...
    'NEJM color scheme','LANCET color scheme','Set1 (Brewermap) color scheme'};

figure;
for m = 1:no_color_schemes
    subplot(no_color_schemes,1,m)
    for k = 1:8
        generatePDF(x(k,:),color_scheme{m}(k,:),'area')
        hold on
    end
    title(titles_schemes{m})
    legend('Data 1','Data 2','Data 3',...
        'Data 4','Data 5','Data 6',...
        'Data 7','Data 8','Location','NorthWest'); legend boxoff
end

alt text


Heatmap

Generating heatmap in MATLAB using the default function heatmap.m (introduced in version 2017a) is quite useful for visualizing the magnitude of elements in matrices. However, the size of the generated heatmaps requires a lot of tweaking to produce a reasonable figure. Here, I have written a wrapper code for making nice appropriate-sized heatmaps quickly and efficiently with minimum input.

Function: figure_heatmap.m

Example 1: Heatmap using only data input

% Generating data
x = randn(10,5);
C = corrcoef(x);

% Heatmap figure
figure_heatmap(C);

alt text

Example 2: Heatmap using a specific color scheme

%Using data of example 1

colorscheme = 'BuGn'; 
%Requires brewermap package
%Download from https://github.com/DrosteEffect/BrewerMap/blob/master/brewermap.m

% Heatmap figure
figure_heatmap(C,colorscheme);

alt text

Example 3: Heatmap with title and labels

%Using data of example 1

colorscheme = 'BuGn'; 
%Requires brewermap package
%Download from https://github.com/DrosteEffect/BrewerMap/blob/master/brewermap.m
text_title = 'Correlation Matrix';
text_labels = {'Variable','Variable'};

% Heatmap figure
figure_heatmap(C,colorscheme,text_title,text_labels);

alt text

Example 4: Heatmap with cell labels and limits of color scheme specified

%Using data of example 1

colorscheme = 'BuGn'; 
%Requires brewermap package
%Download from https://github.com/DrosteEffect/BrewerMap/blob/master/brewermap.m
text_title = 'Correlation Matrix';
text_labels = {'Variable','Variable'};
limits_data = [-1 1]; %for correlation matrix
text_labels_cells{1} = {'A','B','C','D','E'}; %x-axis cell labels
text_labels_cells{2} = {'A','B','C','D','E'}; %y-axis cell labels

% Heatmap figure
figure_heatmap(C,colorscheme,text_title,text_labels,limits_data,text_labels_cells);

alt text

Example 5: Heatmap of a rectangular matrix and automated saving of figure with approproate cell sizes

% Generating rectangular matrix data
X = randn(10,6);
colorscheme = 'YlGnBu';
%Requires brewermap package
%Download from https://github.com/DrosteEffect/BrewerMap/blob/master/brewermap.m
text_title = 'Tall Matrix';
text_labels = {'Variable 1','Variable 2'};
limits_data = [floor(min(X(:))) ceil(max(X(:)))];
text_labels_cells{1} = 1:size(X,2); %x-axis cell labels
text_labels_cells{2} = 1:size(X,1); %y-axis cell labels
savefig = 1;
savefig_name = 'heatmap_example5a';

% Heatmap figure
figure_heatmap(X,colorscheme,text_title,text_labels,limits_data,text_labels_cells,...
    savefig,savefig_name);

%
Y = randn(11,20);
colorscheme = 'BuPu';
%Requires brewermap package
%Download from https://github.com/DrosteEffect/BrewerMap/blob/master/brewermap.m
text_title = 'Fat Matrix';
text_labels = {'Variable 1','Variable 2'};
limits_data = [floor(min(X(:))) ceil(max(X(:)))];
text_labels_cells{1} = 1:size(Y,2); %x-axis cell labels
text_labels_cells{2} = 1:size(Y,1); %y-axis cell labels
savefig = 1;
savefig_name = 'heatmap_example5b';

% Heatmap figure
figure;
figure_heatmap(Y,colorscheme,text_title,text_labels,limits_data,text_labels_cells,...
    savefig,savefig_name);

alt text

alt text

Example 6: Heatmap of a big data matrix

% Generating rectangular matrix data
X = randn(15,50);
colorscheme = 'BuPu';
%Requires brewermap package
%Download from https://github.com/DrosteEffect/BrewerMap/blob/master/brewermap.m
text_title = 'Big Rectangular Matrix';
text_labels = {'Variable 1','Variable 2'};
limits_data = [floor(min(X(:))) ceil(max(X(:)))];
text_labels_cells{1} = 1:size(X,2); %x-axis cell labels
text_labels_cells{2} = 1:size(X,1); %y-axis cell labels
savefig = 1;
savefig_name = 'heatmap_example6';

% Heatmap figure
figure_heatmap(X,colorscheme,text_title,text_labels,limits_data,text_labels_cells,...
    savefig,savefig_name);

alt text


Violinplot

In progress ...

You can’t perform that action at this time.