How to Overlay Curve Fit Over Bar Plot Without Overriding Data?

I'm trying to plot an exponential curve fit on top of this bar chart that I've got, but even when I attempt to use hold on or organize the code a different way it seems to either compress all the data into two separate bins or overrides the data entirely. In essence, I'm trying to predict data three years "into the future", but can't visualize it on the same image.

Here's a simplified version of my code...

Year = 2013:2019;
Video_Cost = [11.09, 17.8, 22.35, 30.06, 37.29, 50.22, 64.83];
Audio_Cost = [3.19, 3.61, 4.75, 6.13, 8.4, 11.54, 14.22];
Total_Cost = Video_Cost + Audio_Cost;

Matrix_Costs = [Video_Cost; Audio_Cost];

figure('Units', 'Normalized', 'Outerposition', [0 0 1 1])
bar(Matrix_Costs.', 'stacked');
hold on
[Curve_Fit, Goodness] = fit(Year', Total_Cost', 'exp1');
xlim([2013, 2022])
hold on
plot(Curve_Fit, 'predobs')

Do I have to use a work around for the bar function here? Or do I need to do some exponential fit using polyfit and polyval with logs? I'm entirely lost so any help would be appreciated.

Answer 1

[Star Strider]

Include the independent variable ‘Year’ in the bar call, and change the xlim values:

Year = 2013:2019;
Video_Cost = [11.09, 17.8, 22.35, 30.06, 37.29, 50.22, 64.83];
Audio_Cost = [3.19, 3.61, 4.75, 6.13, 8.4, 11.54, 14.22];
Total_Cost = Video_Cost + Audio_Cost;

Matrix_Costs = [Video_Cost; Audio_Cost];

figure('Units', 'Normalized', 'Outerposition', [0 0 1 1])
bar(Year, Matrix_Costs.', 'stacked');
hold on
[Curve_Fit, Goodness] = fit(Year', Total_Cost', 'exp1');
xlim([2012.5, 2022])
hold on
plot(Curve_Fit, 'predobs')

See if that does what you want.

.

[Confused Student]

This does solve my issue, but it also introduces new problems with the axis labels, text, and graph title, so I'll have to make a new question for that I guess. Thanks!

[Star Strider]

As always, my pleasure!

None of the text objects were included, so I had no idea there was a problem with them.

.

[Confused Student]

Yeah, I figured it would fix everything up if the two plots became compatible but that didn't happen.

[Star Strider]

I tried to post this reply to your subsequent question, however I guess you solved it in the interim.

I tweaked the text calls a bit.

See if this does what you want —

Year = 2013:2022;
Video_Cost = [11.09, 17.8, 22.35, 30.06, 37.29, 50.22, 64.83, 0, 0, 0];
Audio_Cost = [3.19, 3.61, 4.75, 6.13, 8.4, 11.54, 14.22, 0, 0, 0];
Total_Cost = Video_Cost + Audio_Cost;

Matrix_Costs = [Video_Cost; Audio_Cost];

figure('Units', 'normalized', 'Outerposition', [0 0 1 1])
bar(Year, Matrix_Costs.', 'stacked');
hold on

Exclude_Data = excludedata(Year, Total_Cost, 'Indices', [8, 9, 10]);
[Curve_Fit, Goodness] = fit(Year', Total_Cost', 'exp1', 'Exclude', Exclude_Data);
xlim([2012.5, 2022])
plot(Curve_Fit, 'predobs')
hold on

% for i = 1:length(Video_Cost)

    text(Year, Matrix_Costs(1, :) / 2, compose('\\it %.1f\\rm',Matrix_Costs(2, :)),...
        'HorizontalAlignment', 'center',...
        'VerticalAlignment', 'middle')%, 'Units', 'normalized');
% end

% for i = 1:length(Audio_Cost)
    if (Matrix_Costs(2, :) < 10)
        text(Year, Total_Cost - Audio_Cost / 2, compose('\\it %.1f\\rm',Matrix_Costs(2, :)),...
                'HorizontalAlignment', 'center',...
                'VerticalAlignment', 'middle')

%         t = text(Year(i), Total_Cost(i) - Audio_Cost(i) / 2, compose('\\it %.1f',Matrix_Costs(2, :)),...
%             'HorizontalAlignment', 'center',...
%             'VerticalAlignment', 'middle', 'Units', 'normalized');
    else
        t = text(Year, Total_Cost - Audio_Cost / 2, compose('\\it %.1f\\rm',Matrix_Costs(2, :)),...
            'HorizontalAlignment', 'center',...
            'VerticalAlignment', 'middle');%, 'Units', 'normalized');
    end
% end
% 
% for i = 1:length(Total_Cost)

    t = text(Year, Total_Cost + 1.6, compose('\\it %.1f', Total_Cost),...
        'HorizontalAlignment', 'center',...
        'VerticalAlignment', 'middle');%, 'Units', 'normalized');
% end

set(gca, 'xticklabel',...
    {'{\bf 2013}', '{\bf 2014}', '{\bf 2015}', '{\bf 2016}', '{\bf 2017}',...
    '{\bf 2018}', '{\bf 2019}', '{\bf 2020}', '{\bf 2021}', '{\bf 2022}'})
ylim([0, 195])
xtickangle(45)
ytickformat('usd')

llh = legend('Video', 'Audio', 'Forecasted Trendline', 'Prediction Bounds', 'location', 'eastoutside');
llh.Position(1) = llh.Position(1) + abs(llh.Position(1) * 0.08);

h=gca; h.XAxis.TickLength = [0 0]; h.YAxis.TickLength = [0 0];
h.XRuler.TickLabelGapOffset = 8;
h.YRuler.TickLabelGapOffset = 10;
ylh = ylabel('Total Subscription and Download Cost (USD)');

xlabel('')
tlh = title('Average Annual Expenditure on Streaming Services Per Consumer Unit');

Make appropriate changes to get the result you want.

.

[Confused Student]

Ah yes it does, I appreciate your work!

After looking around for a bit, I noticed that the newer version of Matlab I installed yesterday (to replace the R2015b version I had on my laptop) had nice looking data cursors, so I decided to use them for my project presentation instead for some added visual clarity.

I will definitely use this on the Matlab file I send in for my final submission though.

Thanks!

[Star Strider]

Thank you!

As always, my pleasure!

.