Plotting different sized vectors

I am trying to visualize different participant data from an experiment. Participants took different amounts of time to complete the task and so the length of their time vectors is different. In the experiment, participants indicated feedback to us which we took down a timestamp of. Some participants produced 4 timestamps, some produced as much as 60. At the risk of redundancy, I am going to try to explain this thoroughly: I want to visualize this as a horizontal histogram basically where the horizontal axis is the normalized percentage of their total time and where their feedback occured with repect to their total amount of time they took to complete the task. The vertical axis is the participant number. How can I do this when some produced more data points than others?

Example:

P1: 4, 24, 130, 470, 500

P2, 3, 8, 22, 37, 59, 83, 111, 148, 209

Where the numbers are the timestamps i.e. at 4 seconds, P1 indicated feedback.

Ideally, I would like to make two figures, the first where the horizontal is a time axis that has not been normalized and showcases how some participants took longer than others and the second, described above, where the time axis has been normalized and all points are all plotted against the percentage of the whole time.

[Cris LaPierre]

What count do you want the histogram to be showing? Let's pick the simple case of P1 with 5 times. What should the plot for that participant look like? If I divide all 5 times by the total time, I am left with 5 normalized percentages. Which of these is the value you want to display?

[Emma Walker]

[Emma Walker]

Maybe it should not be a histogram actually. Forget I said that. That is just how I was thinking about it originally. I want all the participants on the same figure.

Answer 1

[the cyclist]

Here is one way. If you have many participants, I would generalize this a bit differently (e.g. by storing participant vectors in a cell array, using a loop). But I don't want to go too far down this path before confirming this is generally what you are looking for.

P1 = [4, 24, 130, 470, 500];
P2 = [3, 8, 22, 37, 59, 83, 111, 148, 209];

figure
hold on
plot(P1/max(P1),1,".","MarkerSize",32,"Color","black");
plot(P2/max(P2),2,".","MarkerSize",32,"Color","black");
set(gca,"YLim",[0.5 2.5],"YTick",[1 2])

[Emma Walker]

This is it, yes! How can I embed this in a loop and do it for all the participants? This is what I am working with.

% Initialize vectors

Timestamps = zeros(24,67); %Hz

TotalTimeVectors = zeros(24,5*1437);

fn = fieldnames(dataStruct); %datastruct contains the timestamps

fn = fn(2:end);

%loop through the participants

figure()

hold on

for i=1: numel(fn)

% Timestamps in seconds to Hz to accomodate oxy-Hb data

% a Timestamp is when a participant speaks an idea

Timestamps(i,:) = 5*dataStruct.(fn{i}); %Hz

TotalTimeVectors(i,1:length( Participants.(fn{i}).data(:,2:2:end) ) ) = 1:1:length( Participants.(fn{i}).data(:,2:2:end) ); %pull HbO, skip Hbr

%insert plotter here

end

[Emma Walker]

How could I use the set function you used for 24 participants?

[Emma Walker]

Here is what I am thinking right now but the figure came out like so:

Timestamps = zeros(24,67); %Hz

%TotalTimeVectors = zeros(24,5*1437);

fn = fieldnames(dataStruct); %datastruct contains the timestamps

fn = fn(2:end);

%loop through the participants

figure()

hold on

for i=1: numel(fn)

% Timestamps in seconds to Hz to accodate oxy-Hb data

% a Timestamp is when a participant speaks an idea

Timestamps = 5*dataStruct.(fn{i}); %Hz

%TotalTimeVectors(i,1:length( Participants.(fn{i}).data(:,2:2:end) ) ) = 1:1:length( Participants.(fn{i}).data(:,2:2:end) ); %pull HbO, skip Hbr

Timestamps = Timestamps(~isnan(Timestamps));

%insert plotter here:

plot(Timestamps/max(Timestamps),1,".","MarkerSize",32,"Color","black");

end

set(gca,"YLim",[0.5 24.5],"YTick",1:1:24)

[the cyclist]

I'm guessing a bit here, because I did not really grok your entire code, but I think it should just be

% plotter
TTV = TotalTimeVectors(i,1:length( Participants.(fn{i}).data(:,2:2:end) ) ); % For convenience
plot(TTV/max(TTV),i,".","MarkerSize",32,"Color","black");

and then

set(gca,"YLim",[0.5 numel(fn)+0.5],"YTick",1:numel(fn))

after the loop.

If that's not quite right, maybe you can figure it out. Otherwise, you probably need to post the data for me to fix it.

[Emma Walker]

I think you were looking at the wrong line. I have removed it for clarity.

I have run the following and have the same result

% Initialize vector

Timestamps = zeros(24,67); %Hz

fn = fieldnames(dataStruct); %datastruct contains the timestamps

fn = fn(2:end);

figure()

hold on

for i=1: numel(fn) %loop through the participants

Timestamps = 5*dataStruct.(fn{i}); %Hz

Timestamps = Timestamps(~isnan(Timestamps));

%insert plotter here:

plot(Timestamps/max(Timestamps),1,".","MarkerSize",32,"Color","black");

end

set(gca,"YLim",[0.5 numel(fn)+0.5],"YTick",1:numel(fn))

So I need to include the y-axis variation in the loop somehow - how do you do that?

[Emma Walker]

Nevermind.. I just realized my blunder. Thank you for your help!

Answer 2

[Steven Lord]

If you have an idea in your mind of roughly what you want the plot to look like, open the Plots tab of the Toolstrip and click the downward facing triangle on the right side of the Plots section of that tab. Find a thumbnail that looks close to what you're envisioning then search for the documentation for the function whose name is below the thumbnail to learn more about how to use it.