posit::conf 2024 was nothing short of amazing! While Posit has already shared their top highlights, we wanted to offer our own take on the experience—what really stood out to us, what we’re excited about, and a few answers to questions that came up during Marcin Dubel’s session.

Let’s get started… 

1. Positron: A Glimpse into the Future of Data Science

One of the biggest buzzes at the conference was about Positron, Posit’s new IDE. It's not just another IDE—it’s designed to handle both R and Python effortlessly, and it’s built on the solid foundation of VS Code.  

What’s really cool is that Positron is focused specifically on data science, not general software engineering. This specialization shows in how it handles different Python virtual environments and multiple R versions seamlessly. 

For those of us who have been working in R, Positron brings some features to Python that we might have taken for granted—like the helper documentation you can pull up with a simple ‘?’ call.

Positron also introduces tools like an R grammar for tree-sitter, which significantly improves GitHub searches for R code. And for those who like to dig deep, Positron’s debugging capabilities even allow you to get inside the C++ code. It’s clear that Positron is going to be a game-changer for data science projects.

We recently shared our thoughts on Positron. Check out this blog post to learn more. 

2. Quarto: More Than Just a Publishing Tool

Quarto was another star of the show, and for good reason. The updates they announced—like new dashboards, Typst, and QuartoLive.

We’ve already had great experiences using Typst in one of our projects. It not only simplified maintenance but also sped up document rendering by about 10x. This kind of performance boost makes a huge difference in projects where time and efficiency are critical.

Quarto also comes with a ton of extensions. Browsing through the community list and official Quarto Extensions, it’s clear that there’s a lot of potential to expand and customize how we use this tool. And with QuartoLive, which brings interactive WebAssembly for documents, we can create documents that are not only interactive with Observable JS but can also be deployed to static servers like GitHub Pages. This opens up new ways to engage our audience and share insights in a more interactive format.

3. Shiny: Getting Even More Interactive

Shiny continues to evolve in exciting ways. Joe Cheng’s demo of an AI-powered sidebot was really something, and Winston Chang’s Shiny Assistant is going to be a game-changer for many of our projects. We’re already thinking about how we can use these tools to build even more interactive and user-friendly applications.

And then there’s the shinywidgets package for Python and the ability to create editable data frames in Shiny—both of which are going to make our lives a lot easier. These updates will definitely help us deliver better, more intuitive applications for our clients.

4. Posit Workbench and Snowflake: Simplifying Data Connections

Another exciting development was the integration of Posit Workbench with Snowflake. Now available through the Snowflake Marketplace, setting up and configuring Posit Workbench is super simplified, running directly from existing Snowflake infrastructure. This integration simplifies our workflow, making it easier to connect and manage data.

The {odbc} package also received some notable updates. It now includes functions for connecting to Databricks and Snowflake, making it simpler than ever to connect to these platforms—especially when a Posit Workbench session is connected to the user’s database accounts. With a Snowflake account connected, each user can see the data according to the permissions set inside Snowflake. This level of integration and security is a big win for our team as we manage large datasets across different platforms.

5. The Art of R Packages: Forging Community with Hex Stickers by Hubert Hałun

We were thrilled to have our very own Hubert Hałun present a talk on something that’s both fun and essential to the R community: hex stickers! Hubert’s session, The Art of R Packages: Forging Community with Hex Stickers, was a hit. He discussed how these little pieces of art are more than just stickers—they’re symbols of community, creativity, and collaboration.

Hubert shared insights on how hex stickers can build a sense of belonging among R users and developers, and how they can be used to foster engagement within the community. It was a unique talk that combined art, culture, and coding, reminding us all that the R community is as much about people as it is about programming.

6. Shiny in Action: Transforming Film Production with TARS by Marcin Dubel

Another highlight for us was Marcin Dubel’s session, Shiny in Action: Transforming Film Production with TARS. Marcin took the audience behind the scenes of a fascinating project where Shiny played a key role in revolutionizing the film production process. TARS is a Shiny-based application that was developed to manage and optimize complex scheduling, resource allocation, and logistics in film production—a process traditionally bogged down by inefficiencies and last-minute changes.

Curious to learn more about how TARS transformed the digital operations of a Fortune 500 movie studio? Read the full case study here.

Marcin’s talk not only showcased the power of Shiny in a completely new industry but also highlighted how adaptable and scalable Shiny apps can be when applied creatively. During the session, Marcin received several interesting questions that couldn’t be addressed live. Here are those questions and Marcin’s responses:

Q1: A single source of data can be great for users but complicated to construct. Did you encounter any resistance when gathering the data to build TARS?
A1: No, as the single data source is in a way a post-product of TARS, not a pre-requirement. TARS serves as a single source of truth from the user's perspective and a home for all new data. Whenever there's a need to include an existing data source, it’s integrated into TARS from the original source, without needing to transfer it. So, there was no resistance, but it was definitely a challenge to get all the connections set up. The challenge was usually more about the organization (finding the right person to grant access) rather than technical issues.

Q2: Did you meet any famous movie stars while working on this project?
A2: Not yet! However, I had the opportunity to visit the studio lot a few times and get much closer to the magic of cinema.

Q3: Who is the audience for these reports? Producers on specific sets? Studio executives? Both?
A3: For the newly created reports in TARS, the beauty of the simple self-report creation process is that anyone can be your audience, depending on your needs. Mostly, these reports are used by department managers to coordinate work between departments.

Q4: If you had a chance to do this project again, what would you change?
A4: There was one assumption about the user's expectations that we had to modify heavily: that each department required the same structure of single-level data (just a list of key-value pairs, fixed for each department). During development, we learned that the needs of each studio department were much more complicated than we assumed. This required us to build a set of different input types, nested values, and dynamic sections. If we were to start the project again, I’d include much more time on the roadmap to handle those complexities.

Q5: Was Shiny as a production-ready tool ever in question? Anything else that was considered?
A5: Not for this project. However, the decision wasn’t purely "from scratch"—we had already worked successfully with the Movie Studio on R Shiny applications for a few years. The internal client team could leverage the existing Posit Connect setup and were used to doing data science with R, so Shiny was a natural fit. However, there was another project for this client where Shiny was questioned—surprisingly, by me, not the client. The application was an internal corporate ChatGPT wrapper in Shiny, supposed to be used by over 10k employees. I suggested using another technology for scalability, but the client convinced me that Shiny would handle it—and it did.

Q6: What error checks did you build into your automated reporting?
A6: During the report generation process, only the data for the scheduled report is snapshotted. If anything goes wrong at this stage, we see an error log in the markdown output. There’s an email notification in Connect if the scheduled report encounters an error. When a user generates the report and encounters an error, there isn’t an automated process. However, there’s a contact form inside the application to get support, and our team can examine the logs to identify and resolve the issue.

6. Posit Connect Cloud: Simplifying Deployment

Posit Connect Cloud was another highlight for us. The idea of turning code from a GitHub repository into a shareable URL with just a few clicks is pretty awesome. We’re definitely going to be using this in our own projects—it’s going to make deploying and sharing our work so much easier.

7. Other Talks We Found Exciting

Aside from our own presentations, there were a couple of other talks that really caught our attention:

• A New Era for Shiny-based Clinical Submissions using WebAssembly by Eric Nantz
  Eric Nantz delivered a compelling talk on how WebAssembly is paving the way for Shiny-based clinical submissions. This is an exciting development, especially for those of us working in life sciences and pharma. The potential for faster, more secure submissions using Shiny and WebAssembly could be a game-changer for regulatory processes.
• Open-Source Initiatives in Pharma - What's Out There and Why You Should Join by Nicholas Masel
  Nicholas Masel’s talk on open-source initiatives in pharma was equally engaging. He highlighted the importance of collaboration in the industry and shared valuable insights on how and why developers should get involved in these initiatives. It was inspiring to see the growing community of open-source contributors in pharma, and we’re excited to explore more opportunities to contribute ourselves.

You can check out the replays of other sessions in “Drugs Not Bugs: Effective Use Of R & Python In Pharma.”

What’s Next for Us

Coming back from posit::conf 2024, we’re buzzing with ideas. This conference has left us feeling inspired and ready to push the boundaries of what’s possible in data science. The new tools and features we learned about are going to shape how we approach our projects. You can find the replays of the sessions after registering on the posit::conf 2024 event portal.

We can’t wait to see where these new tools will take us, and we’re looking forward to continuing the conversation with all the amazing people we met at the conference. A big thank you to the Posit team and everyone who made the conference such a great experience. 

Keep the momentum going! Check out our resources page for valuable insights, practical tools, and tips to boost your projects.