How can I create a heatmap?

Can someone help with to create a heatmap for this data?

I tried with these commands but it didn't work. Thank you!

xvar = T.Xft;

yvar = T.Yft;

Temp = T(:,4:end);

gridres = 100 ;

xs = linspace(min(xvar),max(xvar),gridres) ;

ys = linspace(min(yvar),max(yvar),gridres) ;

[xq,yq]=meshgrid(xs,ys) ;

InterpolatedTemp = griddata(xvar,yvar,Temp,xq,yq) ;

hmap_above = pcolor(xq,yq,InterpolatedTemp);

hmap_above.EdgeColor = [.5 .5 .5] ;

colorbar

colormap jet

title('heatmap')

shading interp

[Chetan Bhavsar]

What Error you are facing?

[Sanley Guerrier]

It says:

Error using griddata

invalid input arguments.

[Sanley Guerrier]

Do you think if I use latitude and longitude instead of using X and Y would give a better heat map plot?

I am not sure what to use, the table below has distance, X and Y coordinates, also latitude and londitude.

[Jon]

Please provide a description of the data in T.xlsx, and what you really want plotted in your "heat map".

Looking at the data, I see that your data X (ft), and Y (ft), do not provide a grid, or even a scattered sampling of the x,y plane. Instead the x and y points all fall along a line. There are multiple columns of T data. If so, the heat map could only plot the data for one of those columns.

At the moment even if you just chose one of those columns, you still couldn't provide a map of temperatures across the x-y plane as you only have data for points along one line in this plane.

[Sanley Guerrier]

Each point (X, Y) coordinate has a series of thermal values from T0 to T6. T0 to T6 are measured a long a straight line which is the width of a road and the column "distance" is the length of the road.

I would like to visualize the data to see the change in temperature a long the road.

Thank you!

Answer 1

[Cris LaPierre]

Do you need to preserve the shape of the road in your heatmap? If not, the easiest approach is to just create a heatmap using the T0-T6 columns.

T = readtable("T.xlsx","VariableNamingRule","preserve");
heatmap(T{:,["T" + (0:6)]})

Another way to do this might be the following

newT = stack(T,["T" + (0:6)],"NewDataVariableName","Value");
heatmap(newT,"Value_Indicator","Distance from CR29 (ft)","ColorVariable","Value")

[Sanley Guerrier]

Thank you. The second way looks more convincing to me. Is there a way to change the color of the map so that the difference between the temperature becomes more obvious.

[Sanley Guerrier]

When I tried these lines of commands the result is different than yours. I don't what's wrong with matlab.

T = readtable("T.xlsx","VariableNamingRule","preserve");

heatmap(T{:,["T" + (0:6)]})

[Cris LaPierre]

Are you using the same dataset you posted here?

Yes, use the colormap command to set your desired colormap.

T = readtable("T.xlsx","VariableNamingRule","preserve");
newT = stack(T,["T" + (0:6)],"NewDataVariableName","Value");
heatmap(newT,"Value_Indicator","Distance from CR29 (ft)","ColorVariable","Value")
colormap jet

[Sanley Guerrier]

It was because I did it in live script. I am so sorry.

[Cris LaPierre]

I am also using a live script. There shouldn't be any difference. If there is, try inserting a figure command before your heatmap command.

[Sanley Guerrier]

I hab multiple live scripts opened and I closed them. Now, it's working fine. Thanks for helping!

[Sanley Guerrier]

Hi Cris LaPiere, last time your help was very useful.

With the same data you helped with last time I wanted to plot a 2D surface with the X , Y coordinates and the temperatures. Could you please help me?

I want to plot all column temperatures in one figure.

N.B: the data I used in the following code is saved in a CSV file.

Thank you!

filename='T.csv';

df = csvread(filename,1,0);%skip 1 header line

X_df=df(:,2);

Y_df=df(:,3);

T0=df(:,6);% when I tried T_df = df(:, 6:end) it gave an error for line 11

%create grid

[x,~,xi]=unique(X_df);

[y,~,yi]=unique(Y_df);

subs=[xi yi];

val=T0;

M=accumarray(subs,val,[],[],NaN);

[X,Y]=ndgrid(x,y);

%create plot

figure(1),clf(1)

surf(X,Y,M)

view(0,90),xlabel('x(ft)'),ylabel('y(ft)'), title("EB Morning Run for T0")

colorbar

shading interp

[Cris LaPierre]

The problem is that you have a 4D data set (X,Y,T,v), or 29x29x7x1. For each (x,y) pair, you have 7 values. A surface plot expects one value per point.

If the heatmap approach worked for you, then I'd do something similar using surf.

T = readtable("T.xlsx","VariableNamingRule","preserve");
surf(categorical(["T"+(0:6)]),T.("Distance from CR29 (ft)"),T{:,["T"+(0:6)]})
colorbar

When you have more than 3 dimensions, you need to get creative, using marker style, color, size, linestyle, etc to convey the higher dimensions.

newT = stack(T,["T" + (0:6)],"NewDataVariableName","Value");
grps = unique(newT.Value_Indicator);
for g = 1:length(grps)
    ind = ismember(newT.Value_Indicator,grps(g));
    scatter3(newT{ind,"X (ft)"},newT{ind,"Y (ft)"},newT.Value(ind),'filled')
    hold on
end
hold off
legend(grps)

[Sanley Guerrier]

Appriciate it Cris!

I think I got now. Thanks for the precision.