IDPLN
Material for synthpop workshop at the IDPLN conference
6th September 2022
Outline for course
All the slides used for the lectures in the course are in this Powerpoint file
RaabNowok_IPDsynthpopLN_synthpop_workshop_060922
The data sets and instructions for each of the practicals are in the table at the bottom of this page
Session 1
Introduction and overview
Lecture 1 Basic synthpop via full conditional models
- Basic ideas
- Parameters you can change – examples
Lecture 2 Utility and tips and tricks for improving it
Practical 1
BREAK refreshments (need to get your own – sorry)
Session 2
Lecture 3 SDC and releasing data
Lecture 4 Other methods
- Included in synthpop
- Not included in synthpop
- Differential privacy
Practical 2
Session | What included |
Instructions and codebook | Data | Sample code |
Practical 1
Quality of Life in Poland 2011 |
Synthesising with conditional models, comparing utility and finding the influence of changing parameters
|
SD2011_tasks | See help for SD2011 in synthpop library | synthesis_SD2011 |
Practical 2 a)
Sample from the American Community Survey 2018. |
Challenge of customising a large data set to maintain all the relationships with cart models, stratification and using catall at the start for difficult variables. | acs2018_tasks.pdf | acs2018.rds | synthesis_acs2018.R |
Practical 2 b)
Internet use in the Scottish Household Survey 2001/2002. |
Utility evaluation, plotting models, defining rules, stratification, catall, structural zeros, differential privacy.
|
shs_tasks | shs.rds | synthesis_shs.R |