Welcome to the DataSHIELD beginners hub!
Please use this page and its sub pages to navigate through the process of understanding the purpose of, installing and using DataSHIELD for the first time.
The structure is in two parts:
Installation, which is specific to the operating system (OS) you are running.
Training in the DataSHIELD environment, which will work on any OS once the installation has been successfully run.
v6.1 Linux Installation Instructions (involves some use of terminal command line)
There are six approachable parts to the training environment of DataSHIELD, designed to last 10-15 minutes each. At the start of every one is a reminder of how to log in (look for the drop down button) in case you wish to start where you left off earlier.
5: Sub-setting
6: Modelling
Where can I find out what data I am looking at? The column names derived from ds.colnames() aren’t very descriptive!
This is what the data dictionary is for! In the tutorial for DataSHIELD above, we connect to the CNSIM dataset. The dictionaries are stored in the Opal server (which, for the tutorial, can be accessed at the IP address http://192.168.56.100:8080) , logging in with the username & password (as in the tutorial), and navigating:
from the homepage, select “Projects” 3rd from left on the top bar;
on the table of projects select the one you are interested in (e.g. CNSIM);
on the next page, select any of the studies in the table (usually all the same parameters) (e.g. CNSIM1);
on the next page the data dictionary is stored with descriptions of each variable.
Why can I not connect to the data after starting my VM?
It often takes 2 minutes for your VM to start, longer if you are loading two VMs simultaneously. It also is dependent on your computer’s processor power and the available RAM. Please be patient if it is being slow, and check back after 5 minutes, your VMs should have powered up successfully and be ready for use!
A ds._function_ has disappeared! DataSHIELD doesn’t recognise it exists! The help for it won’t load!
Try using devtools::check() to force RStudio to recognise it. To do this, you need to have the R Package “Devtools” installed. Instructions for installing it are on the CRAN. The installation does take 15 minutes, beware if you are under time pressure!
N.B. This material is kept up to date for the current version of DataSHIELD release (currently v6.1, see here).