Analytical Data Science Programmer Coding Assessment

by Holger Langkabel

Repo structure

This repo contains one folder for each question in the Analytical Data Science Programmer Coding Assessment. The requested files can be found inside each question folder.

Question 1 (question_1_sdtm)

This folder contains the following files:

• 01_create_ds_domain.R: The requested program
• sdtm_ct.csv: The file for study controlled terminology used by the program
• ds.rds: The final output file of the program run
• 01_create_ds_domain.log: A log generated using the logrx package as evidence of an error-free run

Note on the solution

The expected result dataset is meant to contain the variables VISITNUM and VISIT. However, the mapping instructions in the linked PDF file do not contain any instructions on which variables to map. These variables are neither required as per SDTMIG v3.4. I therefore decided not to create them. In real-life, I would ask for further clarification from the mapping instruction provider or the company data standards group.

Question 2 (question_2_adam)

This folder contains the following files:

• create_adsl.R: The requested program
• adsl.rds: The final output file of the program run
• create_adsl.log: A log generated using the logrx package as evidence of an error-free run

Note on the solution

The specification for LSTAVLDT asks to use ADSL.TRTEDTM. However, this variable does not exist anyplace nor is its derivation specified. In real life, I would have asked for further clarification from the specification provider. For the sake of the exercise, I used ADSL.TRTSDTM.

I would like to point out that - in real life - I would have raised a discussion about the logic behind using AE start date in the derivation of LSTAVLDT instead of AE end date.

Question 3 (question_3_tlg)

This folder contains the following files:

• 01_create_ae_summary_table.R: The requested summary table program
• ae_summary_table.html: The table produced by the above program
• 01_create_ae_summary_table.log: A log generated using the logrx package as evidence of an error-free run