I want to fit a lsline (h=lsline) to a scatterplot in MATLAB. The data is in the xrange between [-2.5 2.5] and I display them with and xlim of [-3 3]. I want to show the lsline only in the range between [-2.9 2.9].
However, when I add h.XData=[-2.9 2.9] the slope of the line changes. Does that make sense to anyone? There is no datapoint between 2.5 and 3 and I expected to only see differences in the length of the line?
Simply change the axis limits after plotting the line:
x = -2.5:.125:2.5;
y = .5 .* x + randn(1,numel(x))*.2;
scatter(x,y);
xlim([-2.9 2.9])
lsline;
xlim([-3 3])
ylim([-1.6 2])
Related
I have a plot in Matlab that shows the intensity as a function of the wavelength (in nm) :
example
I wonder if there is an easy way to color the plot according to the wavelength
like this.
This is the closest I could come up with using only a few lines of code. The colors are probably not precisely matched to the wavelengths, but you should be able to fine tune them by playing with the third (color) argument of the patch command.
x = 350:900;
y = ((x-650).^2)/4000 + 10 + 5*rand(size(x));
patch([x(1) x x(end)], [0 y 0], [0 0 5*(1:numel(x))], 'FaceColor', 'interp')
colormap jet
I'm trying to display a discrete plot with values on the x-axis that are not equally space but I want them to appear equally spaced. I would like a stem plot with the first stick not on the y-axis, and I'd also like to have an horizontal dashed line at y=1.
So far here is what I tried.
x = [10 50 150 3000];
y = [.6 .754 .853 .954];
xv = [1 2 3 4];
stem(xv,y);
set(gca,'XTickLabel',x);
Unfortunately, this is not what I expected. The value on the x-axis are not right and the sticks start on the y-axis and end on the figure edge.
How can I fix this?
EDIT: I initially forgot the horizontal dashed line. Added this.
You just need two tiny additions:
x = [10 50 150 3000];
y = [.6 .754 .853 .954];
xv = [1 2 3 4];
stem(xv, y);
xlim([min(xv)-1 max(xv)+1]); % <--
set(gca, 'xtick', xv); % <--
set(gca, 'xticklabel', x);
You (also) need to explicitly set the xtick option, so that only these ticks are drawn, and no other.
With xlim, you can manipulate the x-axis limits. (Left and right limit might be modified to your needs.)
To add the horizontal dashed line, just add the following at the end:
hold on;
plot([min(xv)-1 max(xv)+1], [1 1], 'k--');
hold off;
(Start and end points of the line might be modified to your needs.)
From Matlab R2018b on, you could also use yline.
The output then looks like this:
When you have a sequence of values that you want to plot equally spaced without any special treatment to what each value actually is, you're essentially defining a set of categories.
MATLAB is good at handling these nicely without any extra trickery to lay them out uniformly on your axes if you declare the values explicitly as categorical.
All you need, therefore, is:
x = [10 50 150 3000];
y = [.6 .754 .853 .954];
stem(categorical(x),y);
yline(1,'--');
ylim([0 1.5]) % Make some space on the y-axis so the horizontal line doesn't sit on the top edge
I would like create a Matlab figure with a double X axis (m/s and km/h) with the same plot.
I have found plotyy and - in Matlab reposity - plotyyy, but I am looking for:
A double X axis.
Together below the plot.
My code is very simple:
stem(M(:, 1) .* 3.6, M(:, 3));
grid on
xlabel('Speed (km/h)');
ylabel('Samples');
M(:, 1) is the speed (in m/s), and M(:, 3) is the data.
I would like only a second line, in the bottom, with the speeds in m/s.
You can do something like the following. In comparison to the solution of #Benoit_11 I do use the normal Matlab labels and refer to both axes with handles so the assignments are explicit.
The following code creates an empty x-axis b with the units m/s with a negligible height. After this, the actual plot is drawn in a second axes a located a bit above the other axes and with units km/h. To plot on a specific axes, insert the axes-handle as the first argument of stem. The conversion from m/s to km/h is directly written in the call to stem. Finally, it's needed to set the xlim-property of the both axes to the same values.
% experimental data
M(:,1) = [ 0, 1, 2, 3, 4, 5];
M(:,3) = [12, 10, 15, 12, 11, 13];
% get bounds
xmaxa = max(M(:,1))*3.6; % km/h
xmaxb = max(M(:,1)); % m/s
figure;
% axis for m/s
b=axes('Position',[.1 .1 .8 1e-12]);
set(b,'Units','normalized');
set(b,'Color','none');
% axis for km/h with stem-plot
a=axes('Position',[.1 .2 .8 .7]);
set(a,'Units','normalized');
stem(a,M(:,1).*3.6, M(:,3));
% set limits and labels
set(a,'xlim',[0 xmaxa]);
set(b,'xlim',[0 xmaxb]);
xlabel(a,'Speed (km/h)')
xlabel(b,'Speed (m/s)')
ylabel(a,'Samples');
title(a,'Double x-axis plot');
As a very simple alternative you could also create a 2nd axis (transparent) and put it below the first one so that you only see the x axis.
Example:
clear
clc
close all
x = 1:10;
x2 = x/3.6;
y = rand(size(x));
hP1 = plot(x,y);
a1Pos = get(gca,'Position');
%// Place axis 2 below the 1st.
ax2 = axes('Position',[a1Pos(1) a1Pos(2)-.05 a1Pos(3) a1Pos(4)],'Color','none','YTick',[],'YTickLabel',[]);
%// Adjust limits
xlim([min(x2(:)) max(x2(:))])
text(2.85,0 ,'m/s','FontSize',14,'Color','r')
text(2.85,.05 ,'km/h','FontSize',14,'Color','r')
Output:
Then you can manually add the x labels for each unit, in different color for example.
The best way i can think to do it is to use 2 plots, for example, you can split the plot into a large and small section by doing something like this:
subplot(100, 1, 1:99) // plot your graph as you normally would
plot(...
subplot(100, 1, 100) // Plot a really small plot to get the axis
plot(...)
b = axis()
axis([b(1:2), 0, 0]) // set the y axis to really small
This is untested, you might need to fiddle around a little but it should hopefully put you on the right track.
I'm using the MATLAB plot feature to compare two vectors. I would like my X axis to represent 1 through 7, and then 14, 21, and then a category at the end for points with undetermined X values..(I'm also not sure how to represent these numberless point (they have Y values, just no X values) I could assign a large number outside any of my X values (1000) to these points, do the 1-7,14,21,1000 and then change the 1000 label to my 'string for un-numbered points'. ??
Here is a way to do it based on the example by The Mathworks here.
The trick is to create 2 x axis with different ranges and make one of them transparent. You can then play around with its properties. I modified a bit their code but kept most of their comments because they explain well the steps.
For the demo I used scatter to represent the points and colored the "good" points in red and the other point (here only 1) in green. You can customize all this of course. For instance I used a value of 100 and not 1000 for the x value of the 2nd axes, but I'll let you figure out how to modify it all to change the output as you wish.
clear
clc
close all
%// Create axes 1 and get its position
hAxes1 = axes;
axes_position = get(hAxes1, 'Position');
%// Create axes 2 and place it at the same position than axes 1
hAxes2 = axes('Position', axes_position);
%// Your data go here.
x = [1 7 14 21 100];
y = rand(1, length(x));
%// Plot the two sections of data on different axes objects
hPlot1 = scatter(hAxes1, x(1:4), y(1:4),40,'r','filled');
hold on
hPlot2 = scatter(hAxes2, x(end), y(end),40,'g','filled');
%// Link the y axis limits and fontsize property of the axes objects
linkaxes([hAxes1 hAxes2], 'y');
linkprop([hAxes1 hAxes2], 'FontSize');
%// Set the x range limits and tick mark positions of the first axes object
set(hAxes1, 'XLim', [1 25], ...
'XTick', [1 7 14 21])
%// Set the x range limits and tick mark positions for the second axes object.
%// Also set the background color to 'none', which makes the background
%// transparent.Add the label for the "super points".
set(hAxes2, 'Color', 'none', ...
'YTickLabel', [], ...
'XLim', [95 100], ...
'XTick', 100,'XTickLabel',{'Super Points'})
And the output:
EDIT
If you wish to add a fit line, I suggest adding a 3rd axes at the same position than the other 2, make it transparent and remove its X- and Y-ticks.
i.e. Add something like this:
hAxes3 = axes('Position', axes_position,'Color','none','YTick',[],'XTick',[]);
And in the call to polyfit/polyval, in my example you want to get only the first 4 elements (i.e. the red ones).
Hence:
p = polyfit(x(1:4),y(1:4),1);
y_poly = polyval(p,x(1:4));
And then plot the line on hAxes3:
hPlot3 = plot(x(1:4),y_poly)
Output:
For this given example:
a=rand(100,1)
ecdf(a)
The obtained plot will have on x proportion of bin values, and on y the proportion of points. I wanted to add a line y=0.5 (50% percent of values) and when cross the line in the plot gather and shows the predicted x value.
It is possible but it suppresses my knowledge.
Thanks in advance
Not sure I interpret your question correctly - but could it be as simple as
a=rand(100,1)
ecdf(a)
hold on
plot([0 1],[1 1]*0.5);
This adds a line at a height of 0.5 from 0 to 1 (which I believe are the limits of the plot that ecdf produced for you).
If you want to find the point where these two lines intersect, you need to obtain the points in the plot using a different form of the ecdf function:
a = rand(100,1)
[f x] = ecdf(a);
figure
plot(x, f); % now you have to make the plot yourself...
hold all
plot(x, 0.5 * ones(size(x))); % add the line at y=0.5
title 'cumulative probability of rand()'
xest = interp1(f, x, 0.5); % interpolate - find x where f would be 0.5
fprintf(1, 'The intercept is at x=%.2f\n", xest);