How to fill the area between two curves with discrete data and with a condition (y<0 & y>0)?

I have a vector containing the values I want to plot in y (data.mat).

In x, I have years ranging from 1979 to 2018.

The data range between negative and positive values.

I want to shade the area between y = 0 and the data when data < 0 with a specific color, and I want to shade the area between y = 0 and data > 0 with a different color.

Since my data is discrete, I cannot fill exactly the region below/above zero since the shift from negative to positive (or vice versa) does not necessarily occur on exact years.

I tried to alleviate this issue by adding 0 to my data at the location where the data was the opposite sign (see commented code for this try). It works better, but the fit is still not exact.

How should I do this to have a region that exactly matches my curve?

Here is what I have tried:

% Data %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
load('data.mat');

year    = [1979:1:2018];            %x axis

y0      = zeros(length(year),1);    % y = 0 vector

index1  = find(data<0);             %index for negative values

%% Figure %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

figure(1)
grid on;hold on;

area(year(index),data(index),'FaceColor', [ 0.7500    0.7500    0.7500]);

%{
%Test (better but still not working)
index2          = find(data>0);             %index for positive values
data2           = data;
data2(index2)   = 0;
Y               = [y0,data2];
area(year,Y,'FaceColor', [0.9000    0.9000    0.9000]);
%}

plt1 = yline(0,'LineWidth',2, 'Color', 'black','LineStyle','--');
plt2 = plot(year,data,'LineWidth',3,'Color','black');

xticks([1979:1:2018]);
ax          = gca;
ax.FontSize = 18; 
xlabel('year','FontSize', 24, 'FontWeight', 'Bold');
ylabel('data','FontSize', 24, 'FontWeight', 'Bold');

Thank you!

Answer 1

[DGM]

If you just want an easy way, consider using this:

https://www.mathworks.com/matlabcentral/fileexchange/61327-anomaly

load('data.mat');
year = [1979:1:2018];
anomaly(year,data);

The colors can be set as you choose, and other plot elements overlaid as before.

[A LL]

Works like a charm, thanks!