Assign each column seperately to each customer.

Dear fellow matlab users

I have stomped upon a problem involving seperating / assigning customers to a column of data. for more explanation read after watching this picture:

i have a set of data looking this ^ .. each column represents 1 customer. so column 1 ==> customer 1. etc.. what i want is make a code that assigns each of the columns to 1 customer in a new array. cause i will divide the 1 column data into 24 (24 hours).. and then plot it.

is it possible to do it using For or while loops..instead of doing it manualy since there is 64 customers.

Regards

a curious matlab user

Answer 1

[Jos (10584)]

Why do you want to separate the columns? To access data of a specific customer, just use indexing into your matrix. This will keep your code very clean. You can retrieve the data of each customer

    for k=1:size(OriginalMatrix,2)
       CustomerData = OriginalMatrix(:,k) ;
       % process, for instance, take the mean:
       MeanValue(k) = mean(CustomerData) ;
    end

but often you can take advantage of ML capability to operate on multiple columns at once! The code above can be replace by the statement:

MeanValue mean(OriginalMatrix,1) ;

[Image Analyst]

I agree with Jos - don't do it. I refer you to the FAQ: http://matlab.wikia.com/wiki/FAQ#How_can_I_create_variables_A1.2C_A2.2C....2CA10_in_a_loop.3F

[awda]

Thanks for the answer :) But: I think i havent expressed my question correctly. what i want the matlab to do is:

i have a 1511x65 double matrix. as picture shows... column 1 is data for 1 customer power usage over 24 hours in 63 days. so 63 x 24 = 1511. which is why its 1511x65..where 65 is the amount of customers.

so i want to make a graph showing one and/or all customers power usage over 24 hours. since it is 63 days envolved..i take the mean of the time for example 3 o'clock of all the 63 days.

i have made a manual code for 1 customer..but is there a way to make it for all of them without the manual way. here is the manual code:

kunde1 = x(:,1);
trow = 24;
tcol = ceil(length(kunde1)/trow);
k1 = [kunde1; nan(tcol*trow-length(kunde1),1)];
k1 = reshape(k1,trow,tcol);
k1(isnan(k1))=0;
  tid = [0:23]';
  plot(tid,mean(k1,2))

NB.: Kunde means customer.

regards a curious matlab user :D

[Image Analyst]

Looks pretty much like Jos's code. Just put it inside a loop over k, and replace your 1 with k to get the kth customer instead of the first customer.

[awda]

unfortunately it is wrong. i want it to show me a graph of all the 64 customers. but 1 graph for each. so i want a for loop using kth customer. where i can call it later to plot with..

right now using jos's code i get wrong data.

since 1 customer have 24x63 matrix.. mean of the 24 hours row will be: 24x1 matrix. and that is good.. if it was possible to make it do that with all of the customer and have a resulting matrix of 24x64.. it would be great.. however with jos code i get wrong mean values even after puting the k instead of the 1 and puting it inside his loop.

any ideas :)