How can I plot multiple graphs from one set of table data, separated by NaN?

I have a table as below: M=

NaN  NaN
2  3
NaN  NaN
NaN  NaN
15  6
6  7
3  12
38  17
9  59
NaN  NaN
10  129
11  34
9  5
NaN  NaN
NaN  NaN
NaN  NaN

Goal: I am trying to plot graphs from each set of data, that is separated by the NaNs

E.g. for the data below a graph with x=2, y=3 Then a graph with x=15, 6, 3, 38, 9, y=6, 7, 12, 17, 59 respectively

Overall goal: to find the slope of each set of data and plot the slopes as a frequency distribution

Thank you!

Answer 1

[Raul Andres Lopez Romero]

You can do this,

first we need to do a logical index for the NaN values,

M= ... %(your data values);
M1=M(1:end,1)
M2=M(2:end,2)
The Index for logical values are (isnan, check documentation for more info)
idxsM1 = ~isnan(M1);
*|*so the index show this*|*
0
 1
 0
 0
 1
 1
 1
 1
 1
 0
 1
 1
 1
 0
 0
 0
M1noNaN=M1(idxsM1)
so you can plot because all NaN values are not in the vector.

plot(M1noNaN)
hold on

plot(M1noNaN-detrend(M1noNaN))

% So detrend (a-detrend(a)) plot the mean value or, slope im not sure check documentation for more info hold off

if you want the vector of detrend sM1=detrend(M1noNaN)

Do the same for vector M2 or second column of you data set M

Answer 2

[Raul Andres Lopez Romero]

I think checking your update, you want something like this

the code for this its simple, you need to generate the second vector without NaNs and do a scatter plot

scatter(M1noNaN,M2noNaN,'filled')
grid on
xticks(0:1:40)
xtickangle(90)
yticks(0:5:140)

and you can check this for the slope

Linear regression

[Shravani Kakarla]

Hi Raul,

Thank you very much!

This helps quite a bit.

My only issue is that defining (end,1) for M1 = M{1:end,1} for each set of data (e.g. 1st data set for the data above is 2,3) is tricky.

Since I have thousands of data sets separated by NaNs, so defining each end by hand is difficult.

Is there any way to parse the data and have an end defined before each new NaN, for each data set?

So M1 would be 2, 3

M2 would be

15 6

6 7

3 12

38 17

9 59

[Raul Andres Lopez Romero]

Check the new answer, then, i think you can manage the data and do all the calculations that you need.

Answer 3

[Raul Andres Lopez Romero]

% Star with position and values
% A it's your matrix that contains the NaN elements 
pos=[true,isnan(A(:,1)).',true];
ini=strfind(pos,[true, false]);
end=strfind(pos,[false, true]) - 1;
M=cell(1,length(ini));
for i = 1:length(ini)
  M{i} = A(ini(i):end(i),:);
end
% That creates a M cell vector to contains all the values 
% But for generate the M1 and M2 ... Mi vectors you need to transform 
%the cell vector in array
M1=cell2mat(M(1,1))
% for M2
M2=cell2mat(M(1,2))
% for M3 
M3=cell2mat(M(1,3))

the result is