There are many possibilities within DataSHIELD. This section of the wiki is organised into bite-size chunks which explain separately how to perform key aspects of analysis. If you prefer full tutorials, an archive of previously given workshops is available here. If you encounter any issues with these tutorials please contact t.j.cadman@umcg.nl.
Two example datashield servers are set up and available to practice on. Each tutorial assumes that you have first logged in to these servers using the steps outlined in Section 2: logging in.
Beginnng to use DataSHIELD has a steep learning curve. We recommend that you start with Section 1: Key concepts where we explain in simple language how analysis with DataSHIELD works.
An Armadillo and Opal server are available to be freely used by anyone learning DataSHIELD. To use the tutorials in the wiki, you first need to login to these servers. Section 2: logging in shows you how. Both servers also contain a number of example datasets which can be used for general practice.
By default DataSHIELD analyses use all datasources; Section 3: datasources shows how to select which datasources are used by a given function.
The tutorials in Section 4: Session management explain useful things you might need to do to manage your session (e.g. loading and saving workspaces), but they do not involve actual analysis.
Section 5 deals with returning non-disclosive information about serverside objects, such as their dimensions.
There are two options for data manipulation in DataSHIELD: (i) Use the functionality within dsBaseClient or, (ii) use dsTidyverseClient. Some tasks can only currently be done with dsBase (e.g. merging data frames), and some researchers may not have access to dsTidyverseClient (e.g. if it is not available within your consortia). However, where functionality is available in dsTidyverseClient and you are able to use this package we recommend it as it is much quicker and more flexible than dsBaseClient.
Section 6.1: Data manipulation with dsBaseClient
Section 6.2: Data manipulation with dsTidyverseClient
Section 7 illustrates how to return descriptive statistics using DataSHIELD and transform them into neater tables.
Section 8: Making figures explains how to make figures using built-in DataSHIELD functions, and how to make your own custom figures using summary statistics.
Many packages are available to conduct analysis in DataSHIELD, and contain their own tutorials about specific analysis. In this section we show how to run two common and useful models: general linear models using both 1-stage virtual pooling and 2-stage meta-analysis. Links to further tutorials are show in the 'further information' section at the end of this page.
Section 9: Linear models with 1-stage and 2-stage meta analyis.
An archive of previous analysis tutorials can be found here.
If you would like to contribute your own tutorials, please read these guidelines.