How to make vertical stripes line contour plot?

Hello everyone,

I have a one dimensional monthly data for 5 years (i.e, 60 points). The data are varying in range between -3 to +3. I want to make a contour plot

of vertical striped lines each of same width corresponding to the value of the each data points.

To clarify more---

in X axis-------there are months.

in Y axis--- Vertical stripe lines of same width and height (may be 4 units). The vertical colour lines intensity shouldd be representing the data value.

first line==(same colour thorught the height). (if data has positive value in this point )

second line==(same color different intensity) (if data has positive value in this point )

So the plot must show the colorbar too.

Any sort of help would be really great.

Thanking you in advance.

Answer 1

[Adam Danz]

dt = datetime(01,01,2000) + calmonths(0:59); % row vector
vals = rand(size(dt))*6-3; % vector (row or col)
plot(dt,vals,'ks')
hold on
dateInt = diff(dt);
dateInt = dateInt([1:end, end])/2;
x = dt + dateInt.*[-1;-1;1;1];
[ymin,ymax] = bounds(vals); 
y = repmat([ymin;ymax;ymax;ymin],1,numel(vals)); 
cvals = 3*ones(numel(vals),1).*-(vals(:)<0);

yline(0)
hold on
patch(x, y, cvals, 'FaceAlpha', 0.25, 'EdgeColor','none')
colormap([0 0 1; 1 0 0])
cb = colorbar(); 
caxis([-3,3])

Alpha scaled to y-values

See "fadescale" to scale the fade.

figure()
dt = datetime(01,01,2000) + calmonths(0:59); % row vector
vals = rand(size(dt))*6-3; % vector (row or col)
vals = sin(linspace(-pi,pi,numel(dt)))*2 + rand(1,numel(dt))*2-1;

% Compute patch verts
dateInt = diff(dt);
dateInt = dateInt([1:end, end])/2;
x = dt + dateInt.*[-1;-1;1;1];
[ymin,ymax] = bounds(vals); 
y = repmat([ymin;ymax;ymax;ymin],1,numel(vals)); 

% Plot, start with underlying patches
figure()
axes()
hold on
h =patch(x, y, vals, 'EdgeColor','none');

% set colormap
nbins = 255;
fadescale = 0.7; % higher is more faded (0:1)
rgb = [ones(ceil(nbins/2),1),linspace(fadescale,1,ceil(nbins/2))' .* [1,1]];
rgb = [rgb; rot90(rgb(1:floor(nbins/2),:),2)];
colormap(flipud(rgb))
cb = colorbar(); 
caxis([-3,3])

% Add data
plot(dt,vals,'ko:','MarkerFaceColor',[.4 .4 .4], 'MarkerSize',4)
axis tight
ylim([-3,3])

[Adam Danz]

You question didn't mention anything about intensity.

I based the solution off of this description:

first line==(same colour thorught the height). (if data has positive value in this point )

second line==(same color different intensity) (if data has positive value in this point )

I interpreted that as "the same color for positive values" and "the same color for negative values" but I think you mean that the y-values should indicate the shade of reds and blues. I'll update my answer.

[Subhodh Sharma]

@Adam Danz Thanks man! But line no 7 there is dimenision mismatch coming, If it is possible Could be pls tell what are you doing in this line. The following error is coming

x = dt + dateInt.*[-1;-1;1;1];
Data inputs must be the same size, or any of them can be a scalar.

Thanks for your help.

[Adam Danz]

What Matlab release are you using?

That line uses implicit expansion which became available in Matlab r2016b. Also, dt must be a row vector.

x = dt + dateInt.*[-1;-1;1;1];

This line creates the x-values for each patch-stripe. If you're using an earlier release of Matlab before r16b, replace that line with this one.

% x = dt + dateInt.*[-1;-1;1;1];
x = dt + [-dateInt; -dateInt; dateInt; dateInt];

[Subhodh Sharma]

@Adam Danz I am using the R2020a. Even there it's showing the same error. Other lines are fine only this line showing issue.

x = dt + [-dateInt; -dateInt; dateInt; dateInt];
Data inputs must be the same size, or any of them can be a scalar.

[Adam Danz]

> I am using the R2020a.

Good. What's the result of

size(dt)
size(dateInt)

[Subhodh Sharma]

Both has the same dimension 1*60

size(dt)
size(dateInt)

ans =

     1    60


ans =

     1    60

all variables

dateInt      1x60              496  duration              
  dt           1x60              480  datetime              
  vals         1x60              480  double  

[Adam Danz]

It turns out that implicit expansion was not supported for duration and datetime values until R2020b (see release notes).

Now I gotta think back to how I did stuff before implicit exapansion....

Try this,

x = [dt-dateInt; dt-dateInt; dt+dateInt; dt+dateInt];

[Subhodh Sharma]

Now this is coming.

patch(x, y, vals, 'EdgeColor','none');
Error using patch
Non-numeric data is not supported in 'patch'

I think I can do from here. Thanks @Adam Danz

[Adam Danz]

Try surf()

Or just update to a newer release ;)

Patch supports datetime objects starting in R2021a (see release notes).

[Adam Danz]

By overlay line do you mean the gray line in my example?

What specific problems are you having? An error? Line not appearing?

Are you plotting the stripes first and then adding the line? If you're plotting the line first, it's under the stripes so you can't see it.

[Subhodh Sharma]

@Adam Danz Thanks. worked