Summary
Linkr is organised into two nested levels. The workspace is a shared container — at the scale of a department, a hospital or a network — that gathers projects, common resources (databases, schemas, wiki, plugins) and a git configuration. The project is the concrete unit of work: a study, a monitoring board, a dashboard, with its data, cohorts, analyses and code.
Two levels of organisation
When you open Linkr, you always move through a two-level hierarchy: you first pick a workspace, then a project inside it. Above them, an organisation (your hospital, university or consortium) acts as a signature: it identifies who produces and publishes the work.
The context
Organisation
The institution behind the work (hospital, university, consortium). An attribution metadata, not a container.
The container
Workspace
Gathers projects and shared resources: databases, schemas, wiki, plugins, git.
The unit of work
Project
A study, a monitoring board, a dashboard — with its data, cohorts, analyses and code.
A workspace contains several projects; a project belongs to a single workspace.
Workspace
Project — Sepsis study
Project — ICU monitoring
Project — COPD cohort
The workspace
A workspace is a shared container. You create one for a collaborative perimeter: most often the team running a health data warehouse or a research network, but just as well a department, a hospital or any other working group. Every member of a workspace sees and reuses the same resources.
A workspace gathers:
Projects
Studies, monitoring boards, dashboards. This is the main content.
Warehouse connections and schemas
The connections to the available warehouses and how to interpret them (OMOP, MIMIC, custom presets), shared across projects.
Cross-cutting resources
Concept mapping, data quality, ETL pipelines, SQL script collections, data catalog — pooled at the workspace scale.
A wiki and plugins
Team documentation and reusable analysis plugins, available in every project of the workspace.
In the app, these resources sit in the workspace sidebar, with the shared Data Warehouse as a group you fold or unfold:
ICU — University Hospital
Data Quality Rule Sets
Create and manage data quality check rule sets for your databases.
OMOP CDM 5.4 — core checks
94%Completeness, plausibility and conformance of the OMOP tables.
MIMIC-IV demo (OMOP)
Vital signs
81%Physiological bounds for ICU vital signs.
MIMIC-IV demo (OMOP)
Unit import
67%Checks on the monthly extract.
ICU extract
In client-only mode (browser only), you create and switch between as many workspaces as you need. What belongs to server mode is attaching people to them: accounts, roles and access rights all assume a backend (see Deployment modes).
A typical setup
A hospital keeps a private workspace (ongoing research, internal dashboards, connected to a private GitLab) and a public workspace (published studies, shared methods, connected to a public repository).
The project
The project is the unit of work. This is where you spend most of your time: exploring data, building a cohort, producing a dataset, composing a dashboard, writing code. A project belongs to a single workspace.
A project brings together everything that contributes to a study or a monitoring board:
- The project warehouse — the linked databases, concepts, cohorts and patient data, read-only.
- The pipeline — the transformations that prepare data for analysis (under construction).
- The lab — the analytical datasets and their analyses, dashboards, reports (under construction).
- The IDE — an integrated Python / R / SQL environment for custom code. It has its own entry, outside the warehouse and the lab.
- Versioning — exporting and sharing the project.
Data quality, concept mapping and ETL pipelines do not appear here: they are set up once in the workspace warehouse, to serve all its projects.
ICU — University Hospital
Cohorts
Cohorts define patient populations from inclusion criteria for your analyses.
Sepsis-3, adults
Unit staySuspected infection and SOFA ≥ 2, first stay.
Septic shock
Unit stayVasopressors and lactate > 2 mmol/L.
Invasive ventilation > 48 h
HospitalizationAdults ventilated for more than two days.
How warehouse, pipeline and lab fit together
These three spaces form the core of a project and follow the path of the data, from raw records to results. That is the topic of the next page: The data pipeline.
A project = an IDE view = an archive = a git repository
A defining principle of Linkr: a project’s content is the same object whichever angle you look at it from. What you see in the IDE, what you get from a ZIP export, and what is versioned in a git repository are one and the same file tree — JSON configuration files, your scripts, your data.
This equivalence has a practical consequence: saving, sharing and versioning a project are the same gesture — handling this file tree. That is what makes ZIP export and git versioning feel natural in Linkr. See Versioning and collaboration.
Further reading
- The data pipeline — how data flows through a project, from the warehouse to the dashboard.
- Versioning and collaboration — sharing and versioning projects and workspaces.
- Your first project — a guided end-to-end walkthrough.