Spatial-Temporal Contour plot from multi excel files
Ältere Kommentare anzeigen
Hello there,
I have four groups of datasets (in excel files, attached). Let’s say they are: data_A1, data_A2, data_A3, data_A4. Each of them contains 3 parameters, namely ax, ay, and az. To make it easily imagined in space-time perspective, let’s say ax is the variable of time and ay is the variable of depth and az is the variable I'd like to plot its spatial-temporal variability. My intention is to get a contour plot of those datasets, representing the spatial-temporal evolution of variable az. So, the contour will plot data ax in x-axis, ay in y-axis, and az value presented by the colormap.
Note that:
- There may be a possibility of overlapping/similar values of ay (depth) between data_A1, data_A2, data_A3, data_A4. In this case, I just need the first record to be considered; or maybe if any other proper handling you could suggest, please welcome.
- There also may be blank values of az.
Thanks!
Akzeptierte Antwort
I am not certain what you want.
As a first approach, this presents them as surfaces —
files = dir('*.xlsx');
for k = 1:numel(files)
filename = files(k).name
T{k} = readtable(filename);
T{k}
[Tstart,Tend] = bounds(T{k}.ax)
dn{k} = datenum(T{k}.ax);
DN{k} = linspace(min(dn{k}), max(dn{k}), numel(dn{k}));
AY{k} = linspace(min(T{k}.ay), max(T{k}.ay), numel(T{k}.ay))
end
for k = 1:numel(files)
Fcn{k} = scatteredInterpolant(dn{k}, T{k}.ay, T{k}.az);
[AX{k},AY{k}] = ndgrid(DN{k},AY{k});
AZ{k} = Fcn{k}(AX{k},AY{k});
end
% figure
% hold on
for k = 1:numel(files)
figure
surfc(AX{k}, AY{k}, AZ{k}, 'EdgeColor','none')
colormap(turbo)
colorbar
xlabel('ax')
ylabel('ay')
zlabel('az')
title(extractBetween(files(k).name,'_','.'))
end
% hold off
It is necessary to use datenum here because scatteredInterpolant does not work with datetime arrays.
If plotted on the same axes, these appear as flat ribbons with respect to ‘az’, and lose their depth (at least in a relative sense). You can of course plot them as contour plots, however I am at a loss as to how to present them in a ‘spatio-temporal perspective’. Animating them will not show here, and it would be difficult to interpolate them over a smooth time scale in any event, in part because their sizes are not the same (although that could be standardised). The other option is to subtract them serially.
.
8 Kommentare
Adi Purwandana
am 18 Okt. 2024
My pleasure!
The sizes can be made the same by defining the same lengths for all the and arrays, for example:
N = 600;
DN{k} = linspace(min(dn{k}), max(dn{k}), N);
AY{k} = linspace(min(T{k}.ay), max(T{k}.ay), N)
My problem is that I do not understand how you want to define ‘spatio-temporal perspective’.
Hypothesizing an approach —
files = dir('*.xlsx');
for k = 1:numel(files)
filename = files(k).name
T{k} = readtable(filename);
% T{k}
[Tstart,Tend] = bounds(T{k}.ax);
dn{k} = datenum(T{k}.ax);
N = 600;
DN{k} = linspace(min(dn{k}), max(dn{k}), N);
AY{k} = linspace(min(T{k}.ay), max(T{k}.ay), N);
end
for k = 1:numel(files)
filename = files(k).name
Fcn{k} = scatteredInterpolant(dn{k}, T{k}.ay, T{k}.az);
[AX{k},AY{k}] = ndgrid(DN{k},AY{k});
AZ{k} = Fcn{k}(AX{k},AY{k});
Check_Sizes = [size(AX{k}); size(AY{k}); size(AZ{k})]
end
%
for k = 1:numel(files)
figure
contourf(AX{k}, AY{k}, AZ{k})
colormap(turbo)
colorbar
Ax = gca;
xt = Ax.XTick;
xtl = datetime(xt, 'ConvertFrom','datenum')
Ax.XTickLabel = compose('%s',xtl);
Ax.XTickLabelRotation = 30;
xlabel('ax')
ylabel('ay')
zlabel('az')
title(extractBetween(files(k).name,'_','.'))
end
figure
hold on
for k = 1:numel(files)
contourf(AX{k}, AY{k}, AZ{k})
colormap(turbo)
colorbar
xlabel('ax')
ylabel('ay')
% zlabel('az')
% title(extractBetween(files(k).name,'_','.'))
end
hold off
axis('padded')
Ax = gca;
xt = Ax.XTick;
xtl = datetime(xt, 'ConvertFrom','datenum')
Ax.XTickLabel = compose('%s',xtl);
Ax.XTickLabelRotation = 30;
figure
hold on
for k = 1:numel(files)
contour(AX{k}, AY{k}, AZ{k})
colormap(turbo)
colorbar
xlabel('ax')
ylabel('ay')
% zlabel('az')
% title(extractBetween(files(k).name,'_','.'))
end
hold off
axis('padded')
Ax = gca;
xt = Ax.XTick;
xtl = datetime(xt, 'ConvertFrom','datenum')
Ax.XTickLabel = compose('%s',xtl);
Ax.XTickLabelRotation = 30;
The contour plots do not overlap because even though the matrix sizes are the same, the ‘ax’ and ‘ay’ variable values differ significantly.
.
Adi Purwandana
am 18 Okt. 2024
Bearbeitet: Adi Purwandana
am 18 Okt. 2024
Walter Roberson
am 18 Okt. 2024
pcolor() is internally surf() and setting the z coordinates to 0 and setting the view to above.
Because it is surf(), it takes the input coordinates as being the coordinates of vertices and it interpolates the face color according the the color of four vertices.
pcolor() is not suitable for interpolating contours.
Adi Purwandana
am 18 Okt. 2024
Bearbeitet: Adi Purwandana
am 18 Okt. 2024
Any other way, the plot of them separately was also looks make sense for me indeed if the range of values is so large difference to put them all into one frame. And I'm just thinking if we can grid the data value of az into cells of ax and ay (for example daily mean of ax), so that properly to look at.
Something like:
I am still not certain what you want, however the heatmap plot gave me the idea to use tiledlayout here.
See if this does what you want —
files = dir('*.xlsx');
for k = 1:numel(files)
filename = files(k).name
T{k} = readtable(filename);
% T{k}
[Tstart,Tend] = bounds(T{k}.ax);
dn{k} = datenum(T{k}.ax);
N = 600;
DN{k} = linspace(min(dn{k}), max(dn{k}), N);
ayv{k} = linspace(min(T{k}.ay), max(T{k}.ay), N);
end
for k = 1:numel(files)
filename = files(k).name;
Fcn{k} = scatteredInterpolant(dn{k}, T{k}.ay, T{k}.az);
[AX{k},AY{k}] = ndgrid(DN{k},ayv{k});
AZ{k} = Fcn{k}(AX{k},AY{k});
% Check_Sizes = [size(AX{k}); size(AY{k}); size(AZ{k})]
end
%
for k = 1:numel(files)
figure
[cm{k},h] = contourf(AX{k}, AY{k}, AZ{k});
LL{k} = h.LevelList
colormap(turbo)
colorbar
Ax = gca;
xt = Ax.XTick;
xtl = datetime(xt, 'ConvertFrom','datenum')
Ax.XTickLabel = compose('%s',xtl);
Ax.XTickLabelRotation = 30;
xlabel('ax')
ylabel('ay')
zlabel('az')
title(extractBetween(files(k).name,'_','.'))
end
figure
tiledlayout(4,1, TileSpacing='none')
for k = 1:numel(files)
nexttile
contourf(AX{k}, AY{k}, AZ{k})
colormap(turbo)
Ax = gca;
Ax.XTickLabel = [];
Ax.YTickLabel = [];
axis('square')
fn = extractBetween(files(k).name,'_','.');
text(min(DN{k}), mean(ayv{k}), sprintf('\\bf%s\\rm: %s ', fn{:}, datetime(mean(DN{k}),ConvertFrom='datenum', Format='dd-MMM-yyyy HH:mm')), 'Horiz','right')
end
The tiledlayout funciton allows for different arrangements of the tiles. See the documentation for that. (Note that if you change them, the text call will also have to change, if you choose to use it.)
.
Adi Purwandana
am 19 Okt. 2024
Bearbeitet: Adi Purwandana
am 19 Okt. 2024
Star Strider
am 19 Okt. 2024
As always, my pleasure!
Weitere Antworten (0)
Kategorien
Mehr zu Discrete Data Plots finden Sie in Hilfe-Center und File Exchange
am 18 Okt. 2024
am 19 Okt. 2024
Community Treasure Hunt
Find the treasures in MATLAB Central and discover how the community can help you!
Start Hunting!
Translated by 
