load_data.m

function [ data, regressionModel, domainBoundaries, predictionVector, theta, varEps ] = load_data(dataSource, nXGrid, nYGrid, offsetPercentage, fitRegressionModel)
%% LOAD_DATA Loads data from various sources
%   Data files are loaded as a function of a variable called dataType
%
%   Input: dataSource
%
%   Output: data, regressionModel, domainBoundaries, predictionVector, theta, varEps
%%
switch dataSource
    
    case 'satellite'
        %% User Input
        % Change as needed
        load('./Data/satelliteData.mat')
        
        % Values of parameters of covariance function [sigma^2, beta, nu]
        theta = [5.57, 0.12, 0.5]; varEps = 0.01;
        
    case 'simulated'
        %% User Input
        % Change as needed
        load('./Data/simulatedData.mat')
        
        % Values of parameters of covariance function [sigma^2, beta, nu]
        theta = [8.13, 0.72, 0.5]; varEps = 0.1;
        
    otherwise
        error('Error. Specified datType is not a valid data set.');
end
disp('Loading data complete');

% Determine the boundaries of the domain spanded by the data.
xmin0 = min(x);
xmax0 = max(x);
ymin0 = min(y);
ymax0 = max(y);
domainBoundaries = [xmin0, xmax0, ymin0, ymax0];

% Make prediction grid
if nXGrid && nYGrid > 0 % If user defines a prediction grid
xPredictionVec = linspace(xmin0 + offsetPercentage, xmax0 - offsetPercentage, nXGrid);
yPredictionVec = linspace(ymin0 + offsetPercentage, ymax0 - offsetPercentage, nYGrid);
[xPredGridLocs, yPredGridLocs] = meshgrid(xPredictionVec, yPredictionVec);
else 
xPredGridLocs = [];
yPredGridLocs = [];
end


% Find observation locations.
logicalInd = ~isnan(values);

% Declare predicition grid
predictionVector = [xPredGridLocs(:),yPredGridLocs(:)];

% Assign lon, lat and observations to data matrix.
data(:,1) = x(logicalInd);
data(:,2) = y(logicalInd);


% Detrend data.
if fitRegressionModel
    data(:,4) = values(logicalInd);
    regressionModel = fitlm(data(:,1:2),data(:,4), 'linear');
    residuals = table2array(regressionModel.Residuals(:, 1));
    data(:,3) = residuals;
else
    regressionModel = nan;
    data(:,3) = values(logicalInd);
end

end