I have a survey that I collect via API in a CSV format (I know it's old school, but that is how they deliver it). Each row consists of the answer from one respondent. Looks like this:
and further in the columns:
I know how to handle surveys with answers and multiple answers, but the number (amount) answers is a puzzle to me.
20 columns consists of amount answers like this:
s_19 header is "How many short conversations with patients?" = 3
s_2 header is "How many short conversations with Familymembers?" = 1
s_20 header is "How many short conversations with Both?" = 1
s_21 header is "How many Long conversations with patients?" = 2
s_22 header is "How many Long conversations with Familymembers?" = 3
s_23 header is "How many Long conversations with Both?" = 1
s_3 header is "How many short conversations with Heartfailure patients?" = 1
s_4 header is "How many Long conversations with Heartfailure patients?" = 1
etc.
Each above is a columns as header and the number is the amount the respondent talked to.
One dimension (target group) of the survey is to show number of Short and long conversations. So I extract that from the data pivotatet as with any other survey (In the above it is s_19,s_2,s_20,s_4 that return short and the other returns long). Another dimension is with who (talk type) (Patient, familymembers, both etc.) - In the above s_19 and s_21 is patients.
Normally when I create a survey in PowerBI I get the data, pivot it and split out the dimensions to count the respondents. But now I also have amount, and one matrix / chart is (target group) and (talk type), but not with the count of respondents, but the amount of people they talked to.
Like this:
How to do that?
I have several columns with numbers but if I think about lowest key value, lets say patient, family etc. I still have to divide between short and long conversations.
Does it make sense?
And how to handle this?
