Experiences from the Research Software Engineering Conference 2023

conference
learning
experiences
software engineering
Some thoughts and experiences learned from the Research Software Engineering Conference 2023 in Swansea, UK.
Author

Richard Ding

Published

October 18, 2023

I attended the RSE Conference 2023 hosted at beautiful Swansea University’s Bay campus and I would love to share my experiences of it in this blog post. This conference illuminated three takeaways for me:

RSE Community: Strengthened and Acknowledged

The RSE Conference 2023 showed that the RSE community is growing up and getting more recognition. At this lively event, software engineers, researchers, and data experts came together. It was a great place for sharing knowledge and making progress in research-related software engineering.

In the last ten years, the RSE community has been figuring out what we do in our jobs, where we fit in, and how people can join us. This helps make research software engineering important for society. Neil Chu Hong’s keynote presentation got us thinking about how our job can keep getting better, bigger, and easier for people who aren’t part of our group. His talk got us talking about how our work affects everyone in society, which was a big theme at the conference.

Responsible Handling of Sensitive Health Data

Imagine you’re a scientist working on important research about healthcare, education, or economics. Some of this research involves really private information, like people’s medical records or financial data. We all agree that this research is vital, but we also have to agree on something equally important: we must keep this sensitive data as safe as we possibly can.

To tackle this challenge, research teams have come up with something called Trusted Research Environments (TREs). These are like special computer spaces designed to be safe and secure, just like a top-secret vault. But they also have to be super useful for researchers, just like a well-equipped lab.

When we talk about sensitive data, we’re talking about information that belongs to real people – people like you and me. We call them “data subjects”. These are the folks whose data is being used for research inside a TRE. We want them to feel safe and comfortable knowing their information is being handled with care.

Then there are the researchers themselves – the scientists and experts doing the work inside TREs. They need a productive space where they can do their research effectively while following all the rules to keep the data safe.

Finally, there are the project leaders or managers. They’re the ones responsible for making sure the research happens safely, responsibly, and that it’s actually getting results.

So, at the RSE Conference 2023, we talked about how to make these TREs safe and productive for everyone involved. We also learned about the Alan Turing Institute’s open-source Data Safe Haven project, which helps with this. During the workshops, we had discussions about what makes a TRE good or bad, what information we’d want to know before using one, and what concerns we have about TREs.

This open discussion allowed us to see how different people – data subjects, researchers, and project leaders – all interact with TREs. It helped us understand the big picture of TREs and how they affect everyone involved. Plus, our discussions will help improve the Alan Turing Institute Data Safe Haven project, making TREs even better in the future.

Python Power: Tools and Applications

The conference was all about Python tools. I got to see lots of cool tools and packages powered by Python that really amazed us.

  • Carrot-CDM: Carrot-CDM is a powerful Python tool designed to simplify the transformation of messy health datasets into the organized OHDSI Common Data Model (CDM) format. Using a command line interface, it extracts input datasets, applies mapping rules defined in a JSON file, and then outputs the transformed data in TSV format, ready for loading into a database or another destination. Carrot-CDM’s clever use of Python classes and a user-friendly approach makes it a valuable ally for research data management, helping researchers make sense of complex health data effortlessly.

  • Apptainer: Apptainer is a Python package that simplifies running complex applications using containers. As an open-source project, Apptainer boasts a welcoming community and stands out in the container landscape for its focus on verifiable reproducibility and security, streamlined integration, ease of mobility with its single-file Singularity Image Format (SIF) container format, and a straightforward yet effective security model. With Apptainer, you work inside a container as you do outside, ensuring a simple yet robust approach to security while harnessing the flexibility of Python-powered containers for your computational needs. Whether you’re a researcher or professional in need of versatile and portable computational environments, Apptainer’s Python-powered simplicity and secure design are your allies in the world of containerized computing.

  • scikit-Learn: scikit-learn is a Python library offers a wealth of data analysis and machine learning tools, but one standout feature is its automated feature selection capability. When diving into complex datasets, it’s often challenging to pinpoint which variables are truly influential for making accurate predictions. scikit-learn simplifies this process by automatically identifying and selecting the most relevant features, eliminating noise, and enhancing the precision of your research models. This feature is a game-changer for researchers, as it streamlines the data exploration process and enables more focused, impactful investigations. So, whether you’re delving into healthcare analytics, economic modelling, or any data-rich domain, scikit-learn’s feature selection empowers you to extract critical insights with surgical precision, elevating the quality and depth of your research findings.

Conclusion: Better Software Better Research

RSEcon 2023 showed us a lot about Research Software Engineering (RSE). We had some interesting talks that got us thinking, and we had sessions that got into the technical details of Python tools and handling data.

As RSEs, we’ve figured out not just what we do but also how and why we do it. We’re responsible for keeping health data safe, and we’ve seen some cool tools like Carrot-CDM and Apptainer that make our work easier. As we go back to our jobs, let’s remember to work together, be innovative, and prioritize ethical and impactful research software engineering. The motto ‘Better Software, Better Research’ isn’t just words; it’s what we aim for.