What can I use a Workspace for?
The BRAINCommons (BC) Advanced Compute Workspace, is a secure, managed AWS cloud computing environment that allows authorised users to access, store and analyse data across multiple datasets using Jupyter Notebooks or R Studio. Using the integrated, web based, Jupyter Notebook environment, users can perform advanced querying and analysis within and across the BC datasets.
Workspaces are aimed at Data Scientists who wish to perform advanced analyses using computational methods in Python or R, therefore good programming and data science skills are highly recommended for all users wishing to use one.
BC Advanced Compute Workspaces can only be accessed by authorised users, through the BRAINCommons environment and only after successful authentication. Workspaces allow for collaboration and can be shared among members of a project team, provided that all project team members have, at least, the same level of access as the owner (Primary Investigator) of the Workspace.
In the context of Workspaces, the following definitions are important to note:
Project: A collection of Data and/or Files. Projects are described by metadata that include, for example, their size, objectives, sponsor, location etc.
Cohort: A set of filters, combining demographic, clinical and project attributes applied to Data, resulting to a recordset of combined Data that the user running the query has permissions to.
Data: Structured data that has been parsed from raw files, curated and mapped to BC data model and can be queried from within a Workspace. Data is stored in Brain Commons database and can be analyzed at the individual data point level.
Files: Unstructured files that contain supplementary information that has not been mapped to BC data model. Case IDs link structured data to unstructured information contained in Files. Examples include docx, xlsx, images (e.g. Dicom files), videos etc. Files can only be queried at the case level (e.g. document containing visit notes for case 12345).
File manifest: A list of unique file IDs and other attributes that can be used within a workspace or via the BC-Client to retrieve the raw, unstructured files for further analysis. File manifests can be created from the Data Explorer page from where they can be exported to a Workspace for use.
Download: The action of copying data or files outside the BC platform. This includes obtaining a copy of the data as JSON, TSV or other structured data format via the User Interface or by using programmatic ways to create copies to cloud services (e.g. S3, Google Drive, FTP sites etc.) through a Workspace or by using the SDK or BC-Client.
Export: The action of pushing a set of case identifiers to a Workspace so that data can be retrieved programmatically in that environment for further analysis. Exporting into a workspace does not create copies of actual data, it creates copies of case identifiers only.
View: The action of viewing Data, Files (e.g. DICOM images), Studies or Charts on screen either through the User Interface of the web application or via a Workspace or Notebook using the SDK.
Save to workspace: The action of storing Data or Files in a workspace directory for further analysis.