The best way to input a table with a unique format

Hi all,

I have a table that has the following format-

I am trying to find the best method to build the table in matlab.

I have tried the following -

systolic=[-inf 119;-inf 119; 120 129;120 129]
diastolic=[-inf 79;-inf 79; 80 84;80 84]  
gender={'Female'; 'Male'; 'Female'; 'Male' }
value= [-3; 0; 0;0]

bpt= table(systolic, diastolic, gender, value)

but I only get the diagonal values.

Any help would be very much appreciated.

Best,

Andrew

[Adam Danz]

What do you mean "I only get the diagonal values."? Here's the result of your code.

[Peter Perkins]

Andew, it seems like maybe what you are looking for is a 3-way table of counts. But it's not really clear.

Also what's not clear is why systoloc and diastolic are each two columns. Two measurements, I guess. And you have made no explanation of value. You are going to need to explain much more about what you have and what you want.

[Andrew Czeizler]

Thankyou! to give a bit more detail.

Basiclly we are asking for user input of the Diastolic and Systolic values and the gender of the patient.

Once the user has entered those values we then find the score on the above table and add it

to a current variable which calculates a final score of risk.

Would the above table suggestion still be the most appropriate?

Could you possibly link me an example?

Many thanks,

Best,

Andrew

[Adam Danz]

The question isn't clear to me. Are you asking wether the format of this table will work with your project?

[Andrew Czeizler]

Hi Adam!

I'm asking the best method to input the table into matlab, so to be able to perform operations on the table.

many thanks

best,

Andrew

Answer 1

[Adam Danz]

(Continuing from comment section under the question)

The best format for a table depends on how you plan on interacting with the table. Basic tidy data principles suggest forming tables that have one row per observation and one column per variable. That makes it really easy to access and interact with single variables. Your current table is organized such the systolic and diastolic columns contain [1 x 2] vectors of [low, high] ranges. You could split that into their own columns like this.

systolic=[-inf 119;-inf 119; 120 129;120 129]
diastolic=[-inf 79;-inf 79; 80 84;80 84]  
systolicLow = systolic(:,1); 
systolicHigh = systolic(:,2); 
diastolicLow = diastolic(:,1); 
diastolicHigh = diastolic(:,2); 
gender={'Female'; 'Male'; 'Female'; 'Male' }
value= [-3; 0; 0;0]
bpt= table(systolicLow, systolicHigh, diastolicLow, diastolicHigh, gender, value)

bpt =
  4×6 table
    systolicLow    systolicHigh    diastolicLow    diastolicHigh     gender     value
    ___________    ____________    ____________    _____________    ________    _____
       -Inf            119             -Inf             79          'Female'     -3  
       -Inf            119             -Inf             79          'Male'        0  
        120            129               80             84          'Female'      0  
        120            129               80             84          'Male'        0  

Now it will be easier to interact with the data. For example, which row classifies the following data?

 syst = 122; 
 dias = 81; 
 gender = 'Female';
 
 rowIdx = bpt.systolicLow < syst & bpt.systolicHigh >= syst & ...
     bpt.diastolicLow < dias & bpt.diastolicHigh >= dias & ...
     strcmpi(bpt.gender, gender); 
 bpt(rowIdx,:)

ans =
  1×6 table
    systolicLow    systolicHigh    diastolicLow    diastolicHigh     gender     value
    ___________    ____________    ____________    _____________    ________    _____
        120            129              80              84          'Female'      0  

bpt.value(rowIdx,:)
ans =
     0

[Andrew Czeizler]

So detailed!

Thank you so much Adam, really clear :).

Best,

Andrew

[Adam Danz]

High five!