Your first research project
What you will achieve. By the end of this tutorial you will have joined your institution's Nuvolos organisation, created a research project, brought in some files and code, invited a collaborator, run a first application against your material, and saved a snapshot you can return to.
How long it takes. About 30–45 minutes, plus the time it takes for an invited collaborator to act on your invitation.
What you need before you start. An invitation email from a Manager at your institution, with the Faculty or Manager role on a Nuvolos organisation. If you have not received one, ask your PI, lab admin, or IT contact - Nuvolos is invitation-only and you cannot create a research project without organisation membership. You should also have one or two files or a code repository in mind to bring in (a CSV, a notebook, a Git URL - anything small).
Step 1 - Accept your organisation invitation
Nuvolos is an invitation-only platform. As a researcher, you can be invited in two main ways: directly to an organisation (the path this tutorial assumes), or to a specific research project as a collaborator (covered separately in the how-to guide on joining a research project).
To create and manage your own research projects you need to be invited to an organisation with the appropriate role:
Member - typically for junior researchers or external collaborators. Does not allow you to create new spaces.
Faculty - the standard role for researchers. Lets you create new research spaces, manage project data, and control access for collaborators.
Manager - has all Faculty permissions, and additionally can manage organisation-wide billing and invite other Faculty/Managers.
For more detail, see Roles and permissions in Reference.
For this tutorial you need either Faculty or Manager. Open the invitation email, click Review Invitation, and complete the sign-up if you do not already have a Nuvolos account. Choose the recommended login option for your institution. Click Accept Invitation. The first sign-up may take up to a minute.
Checkpoint
You should now see your organisation's Dashboard. The Dashboard is where every research project you create or collaborate on will appear.
Step 2 - Create your first research project
A research project in Nuvolos is a space - a workspace where you store and work on your code and data. Each project gets its own space, separate from every other project you have on Nuvolos.
From the Dashboard:
Find the New Research Project button on the Dashboard. If you already have at least three projects, the button is at the bottom of the Recent Projects area.
Click New Research Project.
Create a name and a description. The name has to be unique organisation-wide - try to give descriptive, unique names. You can change both the name and description later.
Choose a visibility option. Private is almost always the right choice and is the safe default for a first project.
(Optional) Add a
README.mdformat description now or later.Click + ADD SPACE.
(Optional) On the next screen, create an application in the Master instance immediately. You can also do this any time later from the Applications screen of the instance - for this tutorial it does not matter which you choose, since Step 4 covers application setup either way.
Navigate directly into the project, or move back to the Dashboard.
Visibility is the only thing you cannot change later
All other properties of a project - name, description, README - can be edited at any time. Visibility is the one exception. If you are unsure, pick Private now; you can revisit only by creating a new project.
Checkpoint
You should now be inside your new project, looking at the Master instance. The Master instance is where you will develop and maintain your research material. You'll learn what role each piece plays in the next steps.
Step 3 - Add some material to the project
Research material is anything that supports the project: data files, source code, notebooks, slides, papers, configuration. There are several ways to bring it in, you can mix and match depending on what you have.
For this tutorial, pick the simplest available path:
Have a small file? Use the file upload feature from the Files view - click UPLOAD on the top left and pick a file.
Have a Git repository? Open a terminal in any application and clone it. If your repository is private, consider adding a Nuvolos-generated SSH key to your accepted keys.
Other options exist too - direct download (wget, curl, or any API from a Nuvolos application), or attaching external storage like S3 buckets via the rclone mount option. For everything available, see Storage and file access.
Pick one method, bring in one piece of material, and confirm it appears in the file list. You can add more later.
Checkpoint
Your file or repository should now be visible in the Master instance's file list. This is the canonical place for project material - everything you do next builds on it.
Step 4 - Run an application against your material
Applications are the interactive tools you actually work in - JupyterLab, RStudio, Matlab, VS Code, and so on. If you skipped the application step in Step 2, install one now:
Confirm in the breadcrumbs that you are in the new project's Master instance, in the Current state.
On the sidebar, click the Screen button (Applications) to see the application list.
Click + ADD NEW APPLICATION. This brings up the Application Marketplace.
Pick something simple to start - JupyterLab if you work in Python, RStudio if you work in R. Click + INSTALL.
Once installed, click the power button or the application name to start it. Wait 30–60 seconds - the first start may take longer if packages or configuration files need to be copied.
When the application opens, you should see the file or repository you brought in during Step 3 in the application's file panel. Open it, run a cell or two if it's a notebook, and confirm the environment behaves as you expect.
Checkpoint
Your application is running, your material is accessible from inside it, and you have run something - even if it's just opening a file or printing "hello" in a cell. This is the moment your Nuvolos project becomes a real working environment.
Step 5 - Invite a collaborator
With material in place and an application running, you are ready to invite a collaborator. Nuvolos supports two kinds of collaborator role; pick based on what you want them to do. To go deeper later, see the full Roles and permissions reference.
Space Administrator - full privileges on the project, including inviting more users, creating new instances, and deleting the project. Use for co-authors who are full partners on the project.
Instance Editor - can see and edit a single instance only. Cannot create instances, invite users, or modify anything outside that instance. Use for collaborators with a narrower scope (a student helping with one analysis, an external advisor reviewing one branch).
For this tutorial, pick whichever fits a real colleague you want on the project. From inside the project:
Click the Cogwheel icon and select Project Users / Instances.
Click + INVITE.
For a co-author with full privileges, choose Option 2 - Administrator Invitation. For a narrower-scope collaborator, choose Option 1 - User Invitation, then pick the instance and the role to assign.
Enter the email address of the person you want to invite. Click + INVITE.
Your collaborator receives an invitation email. Once they accept, they appear in the project users list with the role you assigned.
Step 6 - Save a snapshot
Before you finish for the day, take a snapshot of the Master instance. A snapshot captures the complete state - files, code, application configurations, everything - as an immutable, time-stamped record. It's the cheapest insurance you have.
From the left sidebar, hover the camera icon, and click TAKE SNAPSHOT AND DESCRIBE. Use a descriptive name ("initial setup - material loaded, JupyterLab installed" works well for a first snapshot). Add a short description noting what's in this state.
You're done
You have created a research project, added material, run an application against it, invited a collaborator, and saved a recoverable snapshot. Every later piece of work - exploring alternative approaches, scaling up to a GPU, sharing with a reviewer, exporting the result - builds on these six steps.
Where to go next
To explore alternative methods or parameters in parallel without forking the project, see How-to › Create alternate instances.
To scale up to GPU-enabled compute when your work requires it, see How-to › GPU computation.
When the analysis is complete and you need to package it for a reviewer or hand it off, see second Tutorial: From notebook to reproducible result.
Last updated
Was this helpful?