Scatter plot matrix data while keeping it considered as one data series?

I have two variables, variable A contains information about the orientation of a surface for each minute of a day, variable B contains information about its horizontal tilt. In a perfect world that would mean that they are vectors of size 1440x1. However, at times when the tilt is 0° my program is unable to define a single optimal orientation since all horizontal orientations are equal, so it returns all orientations, making my variables 1440x180 matrices. Where sometimes variable B contains 180 repetitions of tilt = 0° and variable A counts all orientations 1,2,3..180.

I want to scatter plot these so that I have orientation on the X-axis and tilt on the Y-axis. But if I just type:

scatter(A,B);

I get “error using scatter, X and Y must be vectors of the same length.”. Is that because the scatter command doesn’t work with matrices or could it be that I have nans inside the matrices or something else?

If I instead plot it like this:

for i=1:size(A,1)
scatter(A(i,:),B(i,:)); hold on
end

then it doesn’t complain and It plots it like I want but this takes a lot of time and if I turn the legend on I now have 1440 data names since each plot is considered a new set of data. And since it’s considered 1440 individual plots I can’t for example apply a gradient color across the data to represent things that I want.

Does anyone know how to scatter plot a situation like this without using the loop and keeping it one timeseries?

thanks

Answer 1

[Star Strider]

I’m not certain this is an acceptable option for you, but instead of 0° for ‘tilt’, I would substitute it with a very small number (perhaps 1E-8).

[PetterS]

You mean earlier in the process so I end up with vectors instead of matrices? I’m not sure I can do that because I’m getting the orientation/tilt matrices from searching through a 1440x181x180 matrix (minute,tilt,orientation) and having matlab find the largest value for each minute. What I get out from that search are the A and B matrices I described and then I don’t think the situation will change if I adjust the tilt numbers, I already have it in matrix form. And also – having 0° tilt and all orientations is considered a perfectly valid solution for me, so I would still like to have that as an option, I’m just struggling to plot it in a way that makes sense.

[Star Strider]

I misunderstood.

Since scatter only wants vectors, and all your other data are in a vector, I see two options: don’t plot 0° data, or plot them as [0,0]. (My preference would be to not plot the data at 0°, since the orientation there is essentially undefined.)