Source:
vignettes/for_data_cleaner.qmd
Data preparation for Saros
- Ensure that all variables have variable labels (see labelled-package or
{sjLabelled}
).- Item sets (batteries/constructs) sharing the same main question should be meticulously ensured to have the exact same prefix to the label - a slight difference in whitespace will result in them being placed on different graphs.
- Categorical variables must be of type factor or ordered, not the temporary
labelled
class.- Distinguishing between factor-class (unordered) and ordered class matters in Saros - for colour palettes.
- You do not want to create dummy variables.
- Open text variables must be stored as character. These will be visualized/displayed appropriately. Remember that you have the responsibility to consider anonymity!
- Continuous/integer variables are so far not really supported.
- Most other things can be adjusted in the draft_report() argument (see For project managers for more info on configurations).