Thoughputs of each clients are higher than average throughput

Dear community,

I would like to create an OLAP table to calculate throughputs from “Process Start” to"Process End" of each client.
My data model is single eventlog, Case_table, which consists of “Process group”, “Activity”, “Time stamp”, and “Client”.
Every records of case table have client name( client1 - 4) in client column.

I tried to make OLAP table like this.

  • DIMENSIONS
    (1)Cient:
    CASE
    WHEN “Case_table”.“Client” = ‘Client1’ THEN ‘Client-01’
    WHEN “Case_table”.“Client” = ‘Client2’ THEN ‘Client-02’
    WHEN “Case_table”.“Client” = ‘Client3’ THEN ‘Client-03’
    WHEN “Case_table”.“Client” = ‘Client4’ THEN ‘Client-04’
    ELSE ‘その他’
    END

  • KPIS
    (1)Throughput:
    AVG(CALC_THROUGHPUT(CASE_START TO CASE_END, REMAP_TIMESTAMPS(“Case_table”.“Time stamp”, Seconds)))

It looked working, but the result was wired.
The throughput of each client is higher than average throughput.
When I delete dimension, OLAP table shows average troughput lower than each client’s throughput.

Does anyone know how to solve this problem?

Thank you for your support.
Sugano

Dear @Sugano,

try this instead of CALC_THROUGHPUT:

AVG(
    SECOUNDS_BETWEEN(
    PU_FIRST("MAPPING_TABLE", "ACTIVITIES"."EVENTTIME", "ACTIVITIES"."ACTIVITY" = 'CASE_START' ),
    PU_LAST("MAPPING_TABLE", "ACTIVITIES"."EVENTTIME", "ACTIVITIES"."ACTIVITY" = 'CASE_END' )
    )
)

Best regards,
Chanyut

Dear @Chanyut

Thank you for your support.
I tried a following statement as you mentioned above.

AVG( SECONDS_BETWEEN( PU_FIRST(“MAPPING_TABLE”, “ACTIVITIES”.“Time stamp”), PU_LAST(“MAPPING_TABLE”, “ACTIVITIES”.“Time stamp”) ) )

This statement worked same as CALC_THROUGHPUT.
I assume that the dimensions has a problem.

Hi Sugano

In some cases you might have activities between your process path which its thoughput times are lower than avg throughput times. This might lead to an outlier issue.
(Use “Case Explorer” to identify if such cases existed)

Possible Solutions:

  • Use MEDIAN() instead of AVG() to solves statistical outliers issue.
  • Exclude undue activities by using “component filter” for your OLAP

BR,
Chan

1 Like