population inference

[Anonymous]

good afternoon

again here with concerns, in advance I apologize because maybe this is not the best forum for this type of consultation, but as I know in advance the high quality of professionals who are in this place and their good and generous disposition to answer concerns , then I have been encouraged  to ask

I'm really new in the exciting world of statistics, and now I'm doing a corporate job that I consider interesting to solve by chance, but given my little experience I need help to know how to address the situation

we have a help desk in which internal clients are supported, the challenge is to reduce the service ticket through training actions to users and some other knowledge propagation activities. To achieve the above, we requested the analysis of the calls that enter our system. In principle, what you did was a descriptive analysis to know where the largest amount of ticket was grouped, after that classify into categories and calculate the sample size that we should take for each category that we could track.

Now that I have the results of the first category that we inspected and bar graph the types of recurring problems, I get this result

The question is why can I calculate and check the behavior of the total population based on the results of this sample?

This category in particular talks about the problems we have with points of sale, the total number of calls to the help desk was 696, the sample we had to take was 248, although we managed to obtain a total of 257

any help will be appreciated

[Anonymous]

Hi @Anonymous ,

So you means you want to calculate two aggregation result,right? If this is a case, you can consider to write a measure with summarize and Iteration functions to achieve your requirement.

Measure =
SUMX (
    SUMMARIZE ( Table, [Category1], "Count", COUNT ( Table[Column] ) ),
    [Count]
)

If above not help, can you please share some detail information with same data for test?

Regards,

Xiaoxin Sheng

[Anonymous]

first of all thank you very much for coming to my aid @Anonymous , the issue I think is not resolved with the measure.
for this scenario I need to know if I can statistically approximate the total population (696) with the sample that we collected (257), I need to know by what mechanism I can infer the magnitude of cases of the total population and what is the probability that each one will occur from them.
for example the case that is repeated in the sample is "preventive maintenance" = 59
But the question would be if that same distribution can be applied to the entire population.

I appreciate your help in this case

[Anonymous · Accepted solution]

HI @Anonymous ,

Can you please share some sample data with expect result to help us clarify your requirement?

How to Get Your Question Answered Quickly

BTW, you can also add condition filters to summarize function to compare/filter records.

Measure =
SUMX (
    FILTER (
        SUMMARIZE (
            FILTER ( Table, [Column1] = "ABC" && [Column2] <> 123 ),
            [Category1],
            "Count", COUNT ( Table[Column] )
        ),
        [Count] >= 20
    ),
    [Count]
)

Notice: do mask on sensitive data before share.

Regards,

Xiaoxin Sheng