presentations/2020_uRos/notes.md

In matthiasgomolka/sdcLog: Tools for Statistical Disclosure Control in Research Data Centers

What do I do?

• you know my name and where I work
• RDC -> reason why I spent time on an R package for disclosure control in the context of research data centers
• joined work with Tim Becker (also in the audience)
• you need to understand a little bit of what do we at the RDC
• bullets ...
• secure environment

A tiny bit of theory

An example

• ID column, grouping variables, value variables
• calculate grouped mean (using data.table)
• show that result complies to rules
• check

Another example

• same example, now grouping by sector AND year
• problems with both number of distinct entities and dominance

Other functionality

• sdc_model() for all models supported by broom
• sdc_extreme() for the calculation of non-confidential min/max values

Why is it called sdcLog?

• Workflow at our RDC:

  • researchers write their analysees in scripts
  • within these scripts they use the functions mentioned before (name them)
  • researchers source their scripts using sdc_log()
  • sdc_log() creates a log file of everything which happened in the script
  • log file is inspected by RDC staff to evaluate if results comply to RDC rules
  • results can be released

matthiasgomolka/sdcLog documentation built on July 17, 2025, 3:21 a.m.