search

ouhft / COPE

Project Repository for Work Package 4 of the COPE Transplant Trial
https://cope.nds.ox.ac.uk
1 stars 0 forks source link

Make datasets more efficient for Virginia #185

Closed marshalc closed 7 years ago

marshalc commented 7 years ago

From Issue #173:

Exporting to .dta file is an idea but I'm not sure it will necessary saves time, although I have never tried it before. Importing a different format dataset (e.g. csv or xlsx) in Stata is a fairly quick and straightforward process - what really takes time is "setting up" the dataset for analysis once imported, i.e. converting variables from string (text) to numeric type, cleaning data, renaming, labeling, etc. Therefore, if data exported into a .dta file would be the same as those imported from a .csv then it would make no difference. As I said, I have never had data exported directly into .dta files from a database so I don't know what the data would look like so we could always try.

This suggests a discussion with @VirginiaChiocchia could result in creating files in a data structure of Virginia's choosing that allows her to begin analysis much faster, and removes the need to set up a new Stata database each time a data extract is done.

marshalc commented 7 years ago

Park the restructuring the DTS conversation for a later date.

Export the existing data as display labels, rather than database codes.