Open rdstern opened 7 years ago
I don't see any issue with this. It would be good to have something on this in for 0.4.1 since it's for a biometry conference
Here is the first file I would like. This could go into the library into an ILRI directory. I hope someone could work on preparing an RDS file for each of the case studies. One problem is that I would like the variable labels to become part of this information. In the Column metadata pasting is not (yet) permitted. I wonder if that can change.
Given the topic it would be great to get this facility reasonably started by the SUSAN meeting on 20 August.
We are beginning this work today. What are the specific case studies have you worked on? @rdstern
@rdstern
There seem to be 2 sorts of case studies in terms of the data files.
Those with one (or more) Excel files, but each file is simple (i.e. one sheet). a) Those Excel files also seem to have a companion Word file with a description of each column and (sometimes) an overall description. b) The column descriptions can each be pasted into the column metadata. It is sometimes a bit slow, because you can only paste one label at a time. You may also edit the label at the same time, if it is unnecessarily wordy. c) Then the set of Excel files can become a single rds (Instat data) file. d) In those cases the name of the rds file can indicate the case study and can go into the ILRI Biomentics directory.
There are those where the Excel files are more complicated. For now (in those instatnces) I would be inclined (initially) to make each of those a sub-directory and then leave the files as you find them, i.e. for now just copy them over into that directory.
I had that problem. Danny showed that if you click in the label field so you can type into it, then you can paste into that field.
You can't paste multiple fields but that helps.
Yes the single pasting is possible as Roger explains. You don't need to type it in. If it's not clear from the comment how to do that just ask me or Roger and we can show you how to do it. Pasting multiple would be great but we need to do more careful validation which is why it isn't implemented yet.
@mmumbo @PatrykNjoroge this would be really useful for the next version we take to the workshop. Are you able to do more on this this week? Otherwise someone else could take over if you're not available?
We are back already working on them with Patrick for the better part of this week.
That's great.
How is this going? We would like at least some of these files in tomorrow's version.
We are done with all the files at the moment. We will share the files by tomorrow morning since my internet is not stable the moment.
We are done with all the files at the moment. We will share the files by tomorrow morning since my internet is not stable the moment.
This is here https://www.ilri.org/biometrics/default.htm.
It can be downloaded, so it would be easy for us to supply it.
There are 17 case studies and all the data sets are included. When downloaded it is just over 500 mbytes zipped. (About 670Mbytes unzipped.)
The data sets all read easily into R-Instat (at least all I have tried, which is about half the case studies.) This also includes good, though old SSC materials. Some links are not working but it should be easy to get them all to work.
So I suggest we (at least start) for version 4.1.
The one item I am unsure about is how easily R-Instat could cope with quite a long description of a data frame - say 2 or 3 sentences. That is perhaps something to clarify with David and Danny - so I am starting with them as named individuals!