Run your first analysis
The individual pieces of CHORUS are simple enough on their own. What is harder to picture is how they connect. This walks the whole path once, end to end, so the shape of it is clear before you do it with data that matters.
A read-through of the full journey: sign in, open a workspace, start a session, launch a tool, do some work, get the result out. Each step links to the page that covers it properly, so you can come back and go deeper wherever you need to.
What you need first
An account on your institution's instance, and a workspace you have access to. If either is missing, Before you start explains how they come about.
You do not need data of your own to follow this. If your project already has data in a workspace, use that. If not, read it through anyway - the path is the same.
The journey, step by step
Sign in
Open your instance in a browser and sign in. You land on the dashboard, which lists the workspaces you belong to and anything you have running. Nothing is installed on your machine at any point - see Log in & your dashboard.
Open the workspace you will work in
Pick the workspace for your project. Everything from here happens inside it: its data, its tools, its people. If you need one of your own, Create a workspace covers it - though for a real project, the workspace usually exists already.
Start a session
A session is your live working environment inside the workspace. Open one and it starts empty - no tools yet. See What a session is.
Launch a tool
From the App Store, launch the tool you want - Jupyter, RStudio, VS Code, or something specific to your field. It opens as a window in your browser, running inside the workspace rather than on your computer. See Launch an app.
Do the work
Load the project's data and analyse it as you normally would. The data never leaves the workspace, and neither do you - you are simply seeing it streamed to your browser. Several colleagues can be doing the same thing in their own sessions at the same time.
Get the result out
Findings leave through a controlled export, which is reviewed before release. This is the step people forget to plan for, and the one that takes longest - see Exporting results.
Close down, or leave it running
A session keeps running when you close the tab, which is what you want for a long job. When you are finished, stop it - a running session holds resources whether or not anyone is watching. See Resources & lifecycle.
What just happened
Three things were true throughout, and they are the whole design:
The data never moved. You worked on it where it lives, under the rules it lives under. Nothing was copied to your laptop, and there is no second copy anywhere to worry about.
The workspace was the boundary. Your tools, your files, and your colleagues were all inside one isolated environment. Nothing reached out of it and nothing reached in - except through the one door at the end.
That door has a person behind it. The export review is not a formality the platform automates away. Someone accountable decides what leaves.
Why is getting results out the slow part?
It can feel backwards: hours of analysis, then a wait to release a chart.
The reason is that export is the only point where information genuinely leaves the protected environment. Everything else - isolation, access control, audit - is there to make sure that this is the only such point. So it is where the actual judgement happens: does this output reveal anything about individuals, and is releasing it within what the project agreed?
That judgement cannot be automated, and it is not busywork. The practical response is to plan for it: request exports as you go rather than in a rush at the end, and keep them to what you actually need.
Where to go from here
You have seen the path. To go deeper on any part of it:
- Workspaces - isolation, network modes, and working as a team.
- Sessions - what a session is and how it uses resources.
- Apps - the tool catalogue and how versions work.
- Security & governance - the model behind all of it.
If you work with intracerebral EEG, CHORUS.HIP has tutorials for that field specifically.