Mine Çetinkaya-Rundel | Teaching R online with RStudio Cloud

Transcript#

This transcript was generated automatically and may contain errors.

Thank you very much for having me. So today we're going to be talking about teaching R online with RStudio Cloud. And the link to the slides is here on the first slide, and you'll be able to see it on the footer of most of the slides that I'll be going through. That goes to the rendered slides. And I'm not going to be showing much code today or any at all, actually, we're going to be talking about how to set up RStudio Cloud for your teaching. But if you're interested in the source files, I'd be happy to provide a link to that as well.

I think one of the most attractive features is that you, as the instructor, can peek into your students' projects. So we can actually recreate that experience of looking over your students' shoulder, even though you're not in the same room with them.

Getting started: workspaces and projects

When you get to RStudio Cloud, you can get started just like this, and on the login window, you can either create an account or sign up with something like Google or GitHub, or if you have a ShinyApps .io account, that will also work as well. And you'll land into a blank slate like this. So there are two things that you can see here. There's something called your workspace, and I usually like to think about your workspace in the center as the sandbox I play in. I rarely put teaching stuff there, but instead I use separate workspaces for each one of my classrooms.

We will go through each one of these options, but generally, as an instructor, I like keeping my individual workspace as my sandbox for trying things out or one-off sharing, and the individual workspaces for a classroom. So workspaces, when you create an account on RStudio Cloud, you get this workspace of your own. You can add a new workspace and control its permissions. That's why I think it's an attractive option for setting up a class. And projects in either workspace, either the ones that you create specifically for a class or your own, can be public or private.

What is a project in RStudio Cloud? Well, a new project in RStudio Cloud is basically a new project in the RStudio IDE. I've met many people who have been using RStudio for years and years, but don't necessarily use the project features. I'm not necessarily going to be taking a whole lot of time to describe why that's beneficial, but even if you are using RStudio, your local install, I strongly, strongly recommend that you use projects. If you get nothing out of this webinar, but that tip, I think that will still be worth your hour. Trust me, you won't regret it.

And if you do want to learn more about why it is a good idea to be using RStudio projects, even in the local IDE, I have a link to the relevant section in the R for Data Science book in the slides. Similarly, you can create projects from Git, a Git repository. So a new project from a Git repository in RStudio Cloud is the same thing as creating a new project from Git in the RStudio IDE. Both of these actions clone a Git repository. So you have two ways of initiating a project, a blank slate or something that comes from a Git repository.

So we talked about these two options. You have your own workspace or you can create classes. These are your two options for sharing projects with students. We're going to start with the simpler approach, although I will mention why I would recommend the slightly more complicated approach, option two. But option one is you create a single project and you share it with your students. This is how easy it is. It actually, the instructions fit into a tweet. So that's less than 280 characters. Go to RStudio.cloud, log in, create a new project, install any packages, add any scripts or markdown documents that you want to share with your students, change the access level to be available to everyone, copy the URL and share it with your students.

Let's start with creating a new single project. So I'm going and creating a new project. Once the project is deployed, what I can do to start it for, to get it ready for my students is I might, for example, choose to install some packages that I'm going to be using in whatever class example or for the workshop that I am teaching. Any of the packages that you install will then be available for your students as well.

Something I really like to do is just create an R Markdown document quickly. That will actually prompt the RStudio instance to install any necessary packages as well. I usually teach with R Markdown documents, so I usually place a template R Markdown document to get students started with. So if you knit this, the file is going to get saved, which means that when your students may get their own copy of the project, they're going to be able to see that file as well. You might want to rename the project itself so that when you share it, students know, get inherit that name as well. And ultimately, you want to share that URL with your students. But by default, your projects are available only to you. So first you need to change the access level, clicking on the gear icon, and then you can simply send that URL to your students.

Now, let's look at the student view as to how this looks. The student has received that URL from their instructor, they place it into their browser, they log on, and what they're seeing is that same project, you can see with the same name called Workshop Materials, but one thing that is different is that it says Temporary Project on top. So your students need to remember to save a permanent copy. And what is happening at this point is a copy of the project is made for them to further iterate on. If they forget to do that, they lose their changes. But you can see that they're able to get, they got the same document that you had created. And in fact, if we try to just load the Tidyverse package, which we had previously installed on the instructor side, it loads, meaning that the install has actually carried through.

So pros and cons of sharing a single project like this. What's good about it is that students land directly in a project upon login, which means they get the URL from you, and what they see is an RStudio ID, and they can get started. It works well for workshops where all the work that will be done completed is in a single project. But I don't think it works really well where you have, where you're going to have a much longer running class and you're going to have a lot of work that you're going to be doing throughout the class, or if you want to track students in some way. It is also great for sharing code in general. So outside of a teaching setting, if you want to do it for collaboration or to create a reprex , one of the things I love using this feature for is reprexes for shiny apps that tend to be sometimes long, have lots of code, so you can basically put it in an RStudio Cloud instance and share it with somebody else for them to try it out.

The not so good side is that, remember that red reminder on top that said temporary project? Students need to remember to make a copy of the project, and yes, it's red, and yes, it's blinking, but people forget. When I say students need to remember, really what I mean is you as the instructor need to remind them. That places additional burden on you as well, because if they don't, and they work on it, and then they close out of their browser, when they come back, they're not going to be able to see their work.

Another thing that's not so good is that you can't keep track of which students started their assignment. You share the URL with them, and then it just goes out into the ether. You don't have a way of seeing if a student has derived a project from yours, and I'll be able to show you the flip side of this when you use the other paradigm. Additionally, if you are concerned about students sharing work, this is not a great option, because now that a student owns that project, when they make a copy, a non-temporary copy for themselves, they now control the permissions of that project, which means they can simply share it with another friend, and if this is an assignment where you don't want students sharing code, now you've made it additionally easy for them to do so.

Obviously, you controlling permissions of a project does not necessarily mean your students aren't going to cheat if that's what they really want to do, but I think as instructors, it is important for us, if we do want to follow certain rules about sharing work, that we don't use a platform that makes it too easy for students to cheat.

On a remote teaching note, I think one of the things that makes it hardest for sharing a single project is that you can't easily peek into student projects. They would need to explicitly share the project with you. I'm not sure why my bullet point is cut off here, but basically, they need to set the access level to be available to everybody else, and then send the URL to you, as opposed to you being able to track it much more easily. I think that this idea of using RStudio Cloud as a way of peeking over your student's shoulder to help them with any troubles that they're having does not really apply when you share a single project with them.

Option 2: creating a class workspace

The other option is to create a workspace and get your students into that workspace. When I think about creating a workspace, I think about I'm building my classroom, and the people who are invited in here are the people who are in my class, either my students or my helpers or TAs.

How do we go about creating a workspace and then sharing it with our students? As an instructor, I click on new space on the sidebar, give a name for my new space, and also give it a description. This is the name and the description your students will encounter when you ultimately invite them to this workspace. I create the workspace, and then I'm going to do a bunch of prep before I actually invite my students in. That's how easy it is to create the workspace.

The next step is thinking about who is going to be in the workspace. How do I get students into a class workspace? There's a tab called members. If I click on that, you can see that currently, I myself is the only person in this workspace, and my role is noted as admin. On the right-hand side, you can see a bunch of options. Let's drill down into those.

The first thing that you might want to think about is how are you going to invite students into your class? Unfortunately, there isn't just yet an easy way to upload a roster, for example, of student emails. My recommendation would be, instead of choosing invitation require, where you would have to actually individually invite each one of your students, especially if you're teaching a larger class, I would recommend setting the access option to shared, but do so for a short period of time, and share this link with your students. You want to enroll them as contributors, and I will talk about what the other access levels are in a second.

When I say share this link with your students, though, I might recommend, if you are worried about random strangers getting into your workspace, that that sharing is not necessarily done in a public way, so if you keep a public course website, for example, instead of pasting it there, use, for example, your course management software or a roster that you have to email the link to your students. If you don't care about who gets in this workspace at that point, then it doesn't really matter how you share it.

After initial round of enrolment, this could be a drop, or it could be a couple of days if you're having to shift gears mid-semester, switch this access level over to invitation required, and then going forward, you can use invitations for any other people you need to add. Oftentimes, I might have another colleague who wants to see how I'm running things, so I might add them as visitors to my workspace.

The different permission levels, we saw that my name was listed as admin, I am the owner, and I think that role admin really translates to instructor in a teaching setting. I can do everything. I can manage users, view, edit, and manage all projects, in fact, mute the workspace if I want to. Another level is moderator, which in an instructional sense would maybe correspond to a TA. The moderators can view, edit, and manage all projects, but they don't really get to manage users. Contributors are what I would say students. They can create, edit, and manage only their own projects, and a viewer, usually I use it in a course setting as an auditor or a visitor or colleague who might want to see how things are working, they can only view projects that are shared with everyone.

Under the members, that same tab, we also saw these permission levels, and I would recommend unchecking all of these. What are they saying? Contributors can see the members list. This is basically saying, should your students be able to see each other's names? If you think about just your students being able to see each other's names, maybe that's not a big deal, but if you are going to potentially open this course up to the public, you might have concerns about your students' names being publicly available, so I tend to like to uncheck that box.

Contributors can make their projects visible to all members. If this is for a course that is graded where you are expecting that students not share code with each other, let's not make it easier for them to do so, so you can uncheck that box as well. Contributors can change project resources, so you can change how much memory is allocated to certain projects. This might be something you might want to check or you might want to uncheck. I would say, especially if you are teaching introductory students, chances are this is not going to be a concept they are super comfortable with anyway, so you can uncheck that. You can always set that level for them yourself later on. Ultimately, should viewers, like auditors, be able to see the members list? I usually uncheck that too.

The moral of the story here is that you should think carefully before checking these boxes, and if you are not sure, leave them unchecked.

When you create a new workspace, you, by default, can get 10 people in there, including yourself. A lot of us teach courses that have more than 10 students in it, so we have this limit there in order to keep track of usage levels, so you can actually request more members on that members tab, and I would recommend doing so as soon as you set this up. What this means is that if you are setting this up today and wanting to teach it tomorrow morning first thing, that might be a little tough because a human actually needs to approve this, so put in your request as quickly as possible and tell us a bit about what you are using this space for. We are not denying any requests at this point. This is more for tracking of what people are using RStudio Cloud for so that we can try to anticipate what the demand levels might be.

Base projects

Another very attractive feature of RStudio Cloud, if you are using workspaces, is this idea of a base project. What do I mean by a base project? In this base project, the packages you want installed on students' projects live there. Text documents that you want to appear on all student projects also live there. I usually use this for something like a code of conduct, for example, or if you have specific instructions that are the same from assignment to assignment, you can put in things like how do you turn in your work there as well.

The base project applies to all projects created after the base project has been specified. It does not apply retroactively, and that should be an attractive feature. Once people create a project in the class, the settings of that project shouldn't change if you later decide to change how your base project looks. I think most importantly, the base project can be updated as many times as you like throughout the course without having to convince any IT people. There might be a new version of a package that you want your students to learn about, and you can choose to install that package to your base project at 2 a.m. the night before you teach, and tomorrow morning when they get in and start working on an assignment, it will be available to them.

There might be a new version of a package that you want your students to learn about, and you can choose to install that package to your base project at 2 a.m. the night before you teach, and tomorrow morning when they get in and start working on an assignment, it will be available to them.

How does it work to create a base project? There are a few steps for it. You start by creating just like any old project. In this base project, you're going to want to install any packages that you may want to have. I like renaming the project right away as base project. You really can use any name you want, it just needs to be identifiable so that you can choose it from a list later. As I said before, I always like creating an R Markdown document, because that list of packages, I forget what they are. All the list of packages you need to get R Markdown up and running, at least RStudio does it automatically for me, so even though I'm not putting in R Markdown document here, I go through the motions of doing so.

I like adding a plain text file that is basically a code of conduct, so I would add some verbiage to that, but to keep things simple, I'll leave it just with a heading for now. This means this file is going to be copied over to every single project created in this workspace, and then I'm going to want to choose this to be my base project. I'm showing you something here that I almost do every single time, which is that before you can make a project your base project, you need to change its permissions to be available to everyone. We want to change the permissions to be available to everyone, then we click on the gear icon and select the base project by name, and then if at a later point you want to edit it, you can actually do so.

Say tomorrow you decide to teach a topic that uses a package that you haven't used before, you can actually go ahead and add that package to your base project. You can see that once a project has been set as a base project, it disappears from the project list, so now I'm creating a regular project in my class. Let's say this is the homework 1 assignment that I'm creating for my students. I can create a template document for them to start working on, and you will see that as I'm creating this document, RStudio does not prompt me to install the packages that are required for R Markdown, because those already got carried over from my base project. If this is the homework assignment I want to give, I'm going to make it available to everyone, and I'm going to check the box next to make this project an assignment.

Now that I have set up a base project, and the first homework assignment in my class, it is time for me to invite my students. I go to the members tab of my workspace, I choose shared, I say that initial row for my students should be contributors, and I copy the sharing link. It's going to end up being an awfully long link, you might want to use a link shortener or something before you email it to your students, it doesn't really matter. And then your student now receives that email from you with the link to the workspace.

So how do things look like for your student? They go to that link, and this is where they land. They need to log in, and they can use, if they have an existing Google account, log in, it says do you want to join this space, yes, and they land here with your message about this is what this workspace is for, and if they get to the projects, they can see their homework one assignment and start it. So you can see that the R Markdown file you set up for them is there, the document knits without prompting students to install any packages, the code of conduct from the base project is also there as well.