There is anyway that I can plot a graph like this in MatLab with standard deviation ?

[MarKf]

Likely yes.

Not sure tho what you mean with "stadard deviation". As in, you can easily calculate it with stdev = std(data) and you can visualize it in the graph with text (btw, also not sure what Y = 10,657 x 10,222 means), but do you mean SError? do you want the std of the data visualized (like errorbars on the points)? do you want to use std as a variable (like y, and then do a regression on that)? or you want to have the residuals plotted? or do you want the confidence bounds of the regression?

You can do all that and more.

[Vitor]

I wanna errobars on the points with this middle line, but I don t know the function that I have to use

Answer 1

[Image Analyst]

Do you mean like the "errorbar" LinePlot shown about half way down the page here:

https://www.mathworks.com/products/matlab/plot-gallery.html

Answer 2

[MarKf]

n = 15; x = randn(n,1); y = randn(n,1);
sx = std(x); sy = std(y); %maybe non-randn data will have smaller std
lregsrts = regstats(y,x,'linear'); %edit: y first in this function
betal = lregsrts.beta; [X, ixd] = sort(x); %fitlm would be faster, but you can find that example everywhere
Y = ones(size(X))*betal(1) + betal(2)*X; 
scatter(x,y,50,'kd','filled'); %diamonds (romboids, kites) as in OP figure
hold on
plot(X,Y,'k','LineWidth',3); 
errorbar(x,y,sx,'k','horizontal')
errorbar(x,y,sy,'k')
annotation('textbox',[.15 .2 .1 .1] ,'string',...
    ['R^2 = ' num2str(lregsrts.rsquare) newline 's_x = ' num2str(sx)...
    newline 's_y = ' num2str(sy)],'LineStyle','none') %also text
box on, grid on, set(gca, 'YGrid', 'on', 'XGrid', 'off'), hold off% as in OP figure

Edit: This is what was asked but I feel that this is not what they wanted anyway, it's redundant and uncommon. Typical errorbars on a line of data e.g. obtained by averagin (but then not a regression, unless it's not std) it'd be more likely.