aguazul / quiz

Overview

This repository provides data required to complete the tasks described below in a folder called data. These tasks are intended to assess practical programming skills associated with the role of Sr. Data Analyst.

Fork this repository to your own personal copy and develop as needed. Add and commit all source code files you generate. When complete, push your changes to your fork and submit a pull request for review.

Task 1

Develop and document a function to handle "partial dates"

A partial date is a recorded date value that is missing the day or month element, and is commonly encountered in clinical research data. They are typically encoded in a format that is specific to the study and/or the clinical research organization.

In the provided dataset, dates are stored as character strings in YYYY-MM-DD format. Partial dates therein are missing day values which are substituted with "XX", e.g. 2016-01-XX.

In order to process data with partial dates, the missing day values must be filled before use in date comparison operations required for routine reporting.

Goal:

Develop and document an R function to convert partial dates values to values that can be used in date comparison operations.

This function should:

• accept dates as a character vector
• fill in the missing day value with 15, or a user specified value
• return values as objects that can be used in date comparisons

Task 2:

Create a data validation report that flags blood sample collection records missing consent

A critical task of clinical data management is ensuring that the data in a study database complies with data rules specified by the study protocol. In this case patients who have had blood samples collected, should have provided consent on or prior to the date of sample collection. Any records that do not meet this criteria are flagged for review.

There are three tables (provided as CSVs) in the sample data set:

• exam_dates: Stores patient exam visit dates with columns:
  • patient_id: patient id number
  • visit_id: exam visit id number
  • date: date that the visit occurred; may contain partial date
• collection: Stores samples collected from patients
  • patient_id: patient id number
  • visit_id: exam visit id number for sample collection
  • type: sample collection type
  • is_collected: boolean flag if specified sample type was collected
• consents: Stores consents provided by patients
  • patient_id: patient id number
  • category: consent category - e.g. "collection"
  • type: category sub-item for consent
  • date: date that consent was received; provided dates are complete

Goal:

Develop and document an R script that generates a data validation report. This script should perform the following operations:

• Combine (join) the above data tables to align patient_id's with only "blood" sample collection dates and only "blood" sample consent dates.
• Apply the function you developed in Task #1 to process dates as needed.
• Create a column called flag in the combined data set that is TRUE if both of the following conditions are met:
  • a patient blood sample was collected
  • the date of blood sample consent is not on or before the date of blood sample collection
• Export the combined data set to a CSV file called report.csv