When data is labelled in the UI and exported it is only exported with a ,1 set on the labelled data in the csv, data that has not been labelled is left blank, for example:
Due to the fact that the label column has been added, should all unlabelled data be labelled with ,0 rather than just ,?
This would reduce the requirement for the user to have to preprocess the exported csv or deal with NaNs in any upstream processing they did with the exported csv as all the label data would be present.
When data is labelled in the UI and exported it is only exported with a
,1
set on the labelled data in the csv, data that has not been labelled is left blank, for example:Due to the fact that the label column has been added, should all unlabelled data be labelled with
,0
rather than just,
?This would reduce the requirement for the user to have to preprocess the exported csv or deal with NaNs in any upstream processing they did with the exported csv as all the label data would be present.