HEDIMED Research Platform concept aims to create researcher friendly centralized data analytics solution with local like user experience. The system is designed to be scalable and can be deployed with any container supporting environment. The proof of concept deployed in HEDIMED project runs on 5 virtual machines with isolated container runtimes, and the final release is aimed to support single machine setup with fully containerized deployment. The system holds PostgresSQL databases for data storage, Teiid data virtualisation integration layer and Jupyterlab as well as Rstudio (Rocker Project) researcher interfaces.
Instructions
The tool has multiple user roles:
- Data managers import data with common tools for the siloed PostgreSQL databases which are combined to common views with Teiid data virtualisation.
- Researchers can analyse the data with JupterLab or Rstudio web application environments using languages like R, Python or Julia.”
Running the minimal system requires deployment of one or more Postgres data bases on containers, linking them with virtual data bases using containerized Teiid data virtualisation and then creating access to Teiid data interface from containerised researher interfaces (Jupyterlab/Rstudio). By design each researcher should have their personal environment with personal credentials. Access to Research interfaces can be secured with SSH jumphost solution included in the setup. Original data transfer can be handled with provided sftp setup of the system.
The system requires dedicated data and system manager(s) import data with common tools for the siloed PostgreSQL databases which are further combined to common views with Teiid data virtualisation. In principle, researchers can analyse the data with JupterLab or Rstudio web application environments using languages like R, Python or Julia without having a copy of data at all.