Plot data from a table that has years in the rows and months in the columns

17 views (last 30 days)
Maren Hale on 5 May 2021
Commented: Maren Hale on 7 May 2021
I have a dataset in an xlsx file that contains values of Oceanic Nino Index for each month between 1950 and present. The way it's arranged in this file is the years (1950-2021) are in the rows, and the 12 months are in the columns (treat DJF as January, JFM as February, etc.). So each cell of the table has a value for a given month in a given year. See the attachment.
I am trying to make a plot in MATLAB Online where the Oceanic Nino Index is on the y axis and the date (each month of each year) is on the x axis (something like this), but I'm not sure how to plot this when my table is formatted the way it is. Is there a method for making plots when your data is formatted like this, or alternatively is there a method to easily reformat the table to make it easier to plot (i.e. having only two columns, one with year & month and the other with the Oceanic Nino Index value)? Thank you so much.

Scott MacKenzie on 5 May 2021
Edited: Scott MacKenzie on 7 May 2021
Here's an alternative approach that gives a flattened 2D view plus red and blue fills for + and - portions of the graph, as per the example:
yr = T.Year;
y = table2array(T(:,2:end));
y = mean(y,2);
x = 1:size(y);
n = 1000; % interpolate to get smoother +/- transitions
x9 = linspace(1,length(x),n);
y9 = interp1(x,y,x9);
indexPos = y9 .* (y9 >= 0);
indexNeg = y9 .* (y9 <= 0);
area(x9, indexPos, 'facecolor', 'r');
hold on;
area(x9, indexNeg, 'facecolor', 'b');
ax = gca;
ax.XLim = [1 size(yr,1)];
ax.XTick = 1:5:size(yr);
ax.XTickLabels = yr([1:5:size(yr)]);
ax.XLabel.String = 'Year';
ax.XLabel.FontSize = 12;
ax.YLabel.String = 'Oceanic Nino Index';
ax.YLabel.FontSize = 12;
f = gcf;
f.Color = 'w';
f.Units = 'normalized';
f.Position = [.1 .2 .8 .6];
Maren Hale on 7 May 2021
Got it, thanks for the update.

Scott MacKenzie on 5 May 2021
Edited: Scott MacKenzie on 5 May 2021
This might be a reasonable start:
yr = T.Year;
z = table2array(T(:,2:end));
y = 1:size(z,1);
x = 1:size(z,2);
yr = flip(yr); % oldest to newest
z = flip(z,1);
surf(x,y,z);
ax = gca;
ax.ZLim = [-3 3];
ax.XLabel.String = 'Month';
ax.XLabel.FontSize = 12;
ax.YLabel.String = 'Year';
ax.YLabel.FontSize = 12;
ax.ZLabel.String = 'Oceanic Nino Index';
ax.XTick = 1:12;
ax.XTickLabels = { 'Jan' 'Feb' 'Mar' 'Apr' 'May' 'Jun' 'Jul' 'Aug', 'Sep', 'Oct' 'Nov' 'Dec'};
ax.YTick = 1:5:size(yr);
ax.YTickLabels = yr([1:5:size(yr)]);
ax.ZTick = -3:3;
f = gcf;
f.Color = 'w';
f.Units = 'normalized'
f.Position = [.1 .5 .8 .4];
grid on;
Rotated showing a view like in your example:

Eric Sofen on 7 May 2021
Here's another way to tackle this, sticking with using table to contain your data.
% rename the month varnames. just remember that the month represents the
% middle of the 3-month period
t.Properties.VariableNames(2:end) = string(1:12);
% Use stack to go from a "wide" table to a tall time series.
t = stack(t,2:width(t),"NewDataVariableName","Index","IndexVariableName","Month");
t.Time = datetime(t.Year,double(string(t.Month)),ones(size(t.Year)));
plot(t.Time,t.Index)
ylabel("ENSO Index")
% If you wanted to shift to the start of the 3 month period, do
t.Time = t.Time-calmonths(1);
plot(t.Time,t.Index)
I do like Scott's flying carpet plot as well.
Maren Hale on 7 May 2021
Thank you for this alternate approach Eric!

Categories

Find more on Oceanography and Hydrology in Help Center and File Exchange

Community Treasure Hunt

Find the treasures in MATLAB Central and discover how the community can help you!

Start Hunting!

Translated by