Hi! My name is Joanna, and I'm a tableau intern at Fred Hutch Cancer Research Center. I specifically work for HVTN (HIV Vaccine Trials Network.) HVTN runs various clinical trials/studies which we call protocols. There are different clinics around the world in which we perform these studies. These clinics are often called Sites. Each patient is given something called a Ptid (standing for patient ID.) And these patients all have an enroll date (the date they were enrolled (a dimension called Enrolldt)). My project for HVTN is to create tables that give information about the patient enrollment for different sites and protocols. There are two main data points in my tables; the first being the projected/expected number of patients that SHOULD or WILL enroll (the dimension is called Enrollment Projected For Week), and the second is the number of patients actually enrolled (we calculate this by using COUNTD of Patient ID.)
If I create a table including the "Week of Enrolldt" along with a COUNTD of No. (#) Expected Sun Sat (which is the same thing as "Enrollment Projected For Week") and COUNTD of Ptid... then I get a table containing what looks like accurate numbers (Please view Attached Picture #2.) However, if I remove the Enrolldt Dimension, then the Projected Enrollment numbers become obscured and incorrect... but the Number of actually enrolled Patients (COUNTD Ptid) remains accurate(view attached pictures 3, 4, and 5.) Somehow, Tableau is multiplying the actual Projected Enrollment and adding it up to create a really large "Projected Enrollment" that is clearly inaccurate.
Also, if you notice in the 1st attached photo, The Projected enrollment numbers refuse to total correctly. However, to the right, the Distinct count of Patient ID/Ptid Grand Totals correctly. Why isn't Tableau adding and totaling the column of numbers correctly?
The Workbook attached is Tableau desktop version 2018.2