CW24 - Mini-workshops and demo sessions

Session 1: Tuesday, 30 April 2024, 15:45 - 16:45 BST (14:45 - 15:45 UTC)

1.1 Automating assessment of the FAIR Principles for Research Software (FAIR4RS)

Facilitator(s): Kara Moraw (EPCC, University of Edinburgh), Neil Chue Hong (EPCC), Mario Antonioletti (EPCC)

Abstract: A comprehensive set of metrics have been developed by the FAIR IMPACT project to assess the FAIRness (Findability, Accessibility, Interoperability and Reusability) of research software based on the FAIR4RS principles. We extended the data FAIR assessment tool F-UJI (https://github.com/pangaea-data-publisher/fuji) with tests for these software-specific metrics. Some of these tests have been designed to be discipline-specific, as we expect different communities to implement each metric in ways that represent their community norms and standards. In this session, we will demonstrate the tool and focus on a handful of metrics. We will discuss what practical tests for these metrics should look like for different disciplines. Your contributions will shape the implementation of further discipline-specific tests in F-UJI, and after the workshop, you will have a better understanding of how to make your software FAIR.

Audience: You don't need to have any prior knowledge about the FAIR principles, and no technical skills are required. Some experience developing or reusing research software will be helpful for the discussion.

1.2 Sustainable Funding for Open Source Software

Facilitator: Lisa Murphy (Wellcome Trust)

Abstract: The Wellcome Trust has been working to understand how it can most effectively support Open Source Software within its funding programmes and partnerships. We are aware that there are areas of OSS development and maintenance, particularly non-coding elements, which have traditionally been harder to attract funding support for - and thus wish to engage with the community on exploring in more depth what their needs are. In particular, we want to explore how funding can support OSS sustainability, and what the role of a funder might be across areas such as furthering good governance and addressing issues of equality and diversity. Although Wellcome is a funder in the health and life sciences sector we are keen to hear from the OSS community across disciplines, to support our understanding of best practice. We hope that this workshop will help us surface ideas and questions for further exploration, and help us foster connections with the OSS community that we can continue to build on.

Audience: Anyone involved in the creation and maintenance of open source software is welcome to attend, with no preparation required. Although Wellcome funds in the health and life sciences space, we are keen to hear perspectives from across disciplines.

1.3 Community Co-creation

Facilitator(s): Georgia Aitkenhead (The Alan Turing Institute), Bastian Greshake Tzovaras (The Alan Turing Institute), AutSPACEs community members

Abstract: Research and technology are increasingly pushed towards delivering "real world impact", but this push can be a double edged sword: New technology can offer large opportunities for positive change, but also risks embedding and deepening societal issues, for example of access, power and understanding. These risks are particularly pronounced when there is a divide between the people who create technology and research and those who will ultimately be affected by them. In this context, co-creating research and technology with the communities of people who the results are supposed to serve offers one way of responding to this risk. In the context of research software, this can mean engaging directly with a range of people who may not be professional developers themselves, but are the ultimate users or affected by the use of software – to work together to make sure tools are built for and with them. 

In this workshop will explore some of the possibilities of community co-creation, founded on the “nothing about us without us” philosophy of the disability rights movement. This approach is informed by our work with AutSPACEs, a co-created online citizen-science platform designed to investigate sensory processing and autism. Our workshop will kick off by providing some background on the potential and challenges of participatory research co-design using our own work as a point of departure. From this, we will engage in an open-ended, facilitated discussion around two main questions related to co-designing research & technology: 

1. What are the potential benefits of doing co-design in the field of RSE? 

2. What challenges do you expect, or have you experienced, with doing co-design and how could they be addressed? 

Our aim for this session is two-fold: 1. to leave participants with additional knowledge and skills to start thinking about implementing or deepening co-creation in their own work, and 2. to foster bi-directional learning and collective interrogation of the concepts and practice of community co-creation. 

Our session plan itself is the outcome of a co-creation process within our AutSPACEs community and we plan to present the results of it back to our wider community. We will also take collective notes during the discussion sessions, which we will openly publish. This session will be hybrid in order to maximise accessibility. We will also use or experience of facilitating diverse groups to support diverse inclusion.

Audience: There are no prerequisites for attending the workshop. We also do not require participants to have any tools other than the ability to join a video conference call. Our session is open to anyone with any level of experience in community co-creation: from those who have no prior experience or knowledge but are curious to find out more, to those who are experienced practitioners. The open discussion format is designed to allow us to learn from each other as a group.

1.4 Empowering Ethical AI with the Trustworthy and Ethical Assurance Platform: Putting principles into practice throughout the software development lifecycle

Facilitator(s): Kalle Westerling (The Alan Turing Institute), Chris Burr (The Alan Turing Institute), Sophie Arana (The Alan Turing Institute), James Byrne (British Antarctic Survey), Marina Strocchi (Imperial College London), José Alonso Solis Lemus (Imperial College London)

Abstract: The Trustworthy and Ethical Assurance (TEA) platform is an innovative, open-source tool developed by researchers at The Alan Turing Institute and University of York. It is designed to facilitate the creation, management, and sharing of assurance cases for AI/ML tools in science and innovation. Overall, the platform aims to make the complex world of assurance accessible to all stakeholders involved in the development and deployment of data-driven technologies. 

The workshop is designed to offer a deep dive into the TEA platform's ability to support the iterative development of assurance cases. These cases are foundational, providing structured, evidence-based justifications that AI/ML tools are meeting the imperative benchmarks of ethics, safety, and sustainability. You will engage with the platform and learn how it fosters enhanced transparency and accountability in AI/ML projects and serves as a critical bridge fostering collaboration across various disciplines. This kind of collaboration is essential for ensuring that ethical considerations are front and center in the development and deployment of AI/ML tools. 

Moreover, our session will showcase how the TEA platform is instrumental in guiding stakeholders through the ethical landscape of AI/ML, aiding them in making informed decisions that uphold ethical standards. This guidance is crucial for building trust among users, funders, and the wider public, as it demonstrates a committed adherence to ethical principles and safety standards. 

You will learn about the platform's features and capabilities, get live demonstrations and guided skills and training activities, and see a real-world case study in action. In the real-world case study, an assurance case is created within a cardiac research software platform, CemrgApp, focusing on the core attributes of bias mitigation, diversity & inclusivity, non-discrimination, and equitable impact of the system. The focus of the case study is to understand how to set the scope of an assurance case and build the claims and evidence necessary to address the selected core attributes. James Byrne, Research Software Engineer with the British Antarctic Survey, will also speak to the importance of assurance in environmental data science applications in future Digital Twins. 

Through these steps, you will discover how the platform advocates for continuous learning and adaptation, encouraging you to refine and improve your AI/ML tools continually. 

By the end of the session, you will be equipped with the knowledge and tools necessary to apply and implement trustworthy and ethical assurance practices in your AI/ML projects. By integrating these practices, you will be contributing to a more sustainable and ethically grounded research software ecosystem, fostering innovation that respects both ethical considerations and long-term societal objectives.

Audience: This workshop is designed for researchers, software developers, project managers, and policy makers interested in incorporating trustworthy and ethical assurance practices into the design, development, and deployment of AI/ML tools for science and innovation. The session will cater to both in-person and remote attendees, with the software hosted in the cloud to ensure accessibility for all. We are expecting no more than 20 participants to facilitate an interactive and engaging learning experience. 

Participants are not required to have prior experience with assurance case development, or practical knowledge of AI/ML system development and engineering. However, they should have a good conceptual understanding of AI/ML.

Session 2: Wednesday, 1 May 14:40 - 15:40 BST, (13:40 - 14:40 UTC)

2.1 Celebrating 5 years with The Turing Way

Facilitator(s): Kirstie Whitaker, Anne Lee Steele, Emma Karoune, Alexandra Araujo Alvarez (all The Alan Turing Institute)

Abstract: In 2019, at the Collaborations Workshop in Loughborough, The Turing Way (https://the-turing-way.netlify.app/index.html) was launched as a guide to reproducibility, providing tools, methods, and practices to address the reproducibility crisis in science. When we received a second wave of funding from The Alan Turing Institute we reflected further on what skills are required to maximise efficiency, reusability and openness in data intensive science. The Turing Way added guides for project design, ethics, communication and collaboration in 2020. What began as an open-source project building a handbook for data science, has blossomed into a dynamic, global collaborative community.

With 450+ co-authors who have shaped its narrative so far, The Turing Way has unfolded into a multi-guide book featuring 300+ chapters. In this session we will highlight three chapters in line with the themes of Collaborations Workshop 2024: 

• Code Testing (https://the-turing-way.netlify.app/reproducible-research/testing): We will revisit one of the first chapters we published with a particular focus on making sure that there is a human-in-the-loop when using AI tools for software engineering. Imagine a world where the same large language model writes the code and the tests, marking its own homework with no critical oversight!
• The Environmental Impact of Digital Research (https://the-turing-way.netlify.app/ethical-research/activism/activism-env-impact.html): We will highlight how The Turing Way has inspired and joined forces with other communities including Pangeo (https://pangeo.io) and the Environmental Data Science Book (https://edsbook.org/welcome.html) to contextualise open research practices for environmental data scientists, as well as considering and minimising the environmental impacts of data science. 
• Running inclusive events for research incorporating patient and public involvement and engagement (https://github.com/the-turing-way/the-turing-way/pull/3550). We will aim for a 5th birthday live merge of a recent chapter written in collaboration with the AI for Multiple Long Term Conditions Research Support Facility (https://www.turing.ac.uk/research/research-projects/ai-multiple-long-term-conditions-research-support-facility) and The Turing Way Accessibility working group as we continue to break down the barrier between “citizen” and “scientist”. 

After these lightning talks we will have small discussion groups focused on what we have achieved in the last 5 years and what we predict will change in the next 5 years. At the end of the session we will sing happy birthday to our community and provide cake for anyone who is attending in person.

Audience: There are no prerequisites for joining this session. In fact, welcoming questions from people who are not part of the existing Turing Way community is one of our primary aims. We will use slido for questions and polls, and miro and post it notes for collaboration online and in person respectively. We hope that some of the existing community members and the “older” reproducibility and open science experts will join us for a slice of in-person or virtual cake and to reminisce on how far we have come (and how far we have to go).

2.2 Pin the tail on the repo: Why predicting OSS sustainability based on Git repos is merely a stab in the dark

Facilitator(s): Yo Yehudi (OLS / University of Manchester)

Abstract: "Pin the tail on the donkey" is a children's game, where the goal is to successfully complete a goal whilst disoriented and unable to observe the full context of the challenge due to a blindfold. 

This mini-workshop is inspired by the game, but instead will ask participants to examine characteristics of Git repositories (such as number of committers, time to close pull requests, presence of certain files such as Code of Conduct, readme, contributor guidelines) and attempt to "pin the tail" by predicting the age of each repo based on its characteristics. The goal of this exercise is to (hopefully) demonstrate that Git and Github-based metrics alone can not effectively provide enough context to provide sustainability predictions. 

For in-person attendees, we'll also try to bring a paper donkey so you can try the original game :) 

Audience: This will be suitable for anyone, online, or in-person, and will not require any pre-preparation from participants :).

2.3 Indoor air quality in schools: the SAMHE citizen science project

Facilitator(s): Sam Bland (SEI York, University of York)

Abstract: Schools’ Air quality Monitoring for Health and Education (SAMHE) project is a UK-wide citizen science project which measures indoor air quality in UK schools and provides them with access to a Web App. In this app, teachers and pupils can observe and engage with their monitor data. Pupils can also perform activities and experiments to develop their own knowledge about air quality whilst collecting contextual data for researchers to analyse. 

Throughout the project we faced challenges of ethics, project management, technical hurdles and content creation. A Co design approach was used to scope the project and ensure that we hit the needs of both schools and researchers. The result was a complex set of requirements and user stories defining the need for data flow from both the schools to the researchers and also from the researchers to the schools. We used many project management and technical solutions to overcome the issues including engaging all researchers in the app design and testing which was managed through github projects. We also built tools to allow researchers to directly create content for the schools reducing the burden on software developers. 

This session will begin with an introduction to the SAMHE project covering the challenges and solutions of working on a citizen science project that connected a multidisciplinary team of researchers with over 1000 schools across the UK. Demonstrations of the app will be given showing how it is used by school students, and there will be opportunities to give feedback and suggestions for improvement. 

The second half of the workshop will be a design jam to design a new activity for the app which could be implemented and delivered to schools. This activity will use Miro(or similar) to engage attendees with the complex challenges of designing an activity that meets the needs of a broad age range of students, provides valuable data to researchers and above all is fun! We will provide guidance and facilitate discussion on some of the ethical, technical and data challenges that could be faced in the implementation of their activity. 

Audience: This session will be of interest to anyone working with schools, web app development, co-design or air quality.

2.4 Best Practices in Green Software: Principles, Patterns and Tools

Facilitator(s): Colin Sauze (National Oceanography Centre), Sadie Bartholomew (NCAS/University of Reading)

Abstract: There are growing concerns over the climate impact of computing and how we design and build software is a big part of this. This demo session will cover some of the material created by the Green Software Foundation to reduce the impact of software. It is aimed at people who develop or commission the development of research software such as Research Software Engineers and Researchers and their manager(s). This session will be in three parts: 

• An introduction to some high level principles of green software and how to calculate the emissions of a piece of software using the Software Carbon Intensity metric. 
• An overview of some specific design patterns intended to reduce emissions. 
• A demonstration and a chance to test out some key tools that can help with measurements including the Carbon Aware SDK, Firefox Power Profiler and CO2.js. 

Many of the Green Software Foundation tools are aimed at large websites running on cloud providers. We will focus on the aspects that are most relevant to the research software community but will also be seeking feedback from the audience about what is most relevant to them with the aim of building some research software specific guidelines as a follow on activity.

Audience: This would suit anyone who uses or develops software and wants to educate themselves about reducing its environmental impact reduction. Some general familiarity with the process of software development is needed. Programming experience would be helpful but isn’t essential. To test out the Firefox Power Profiler yourself you will need Firefox installed and if running Windows it may need to be Windows 11. To test out CO2.js you will need to be able to install and run NodeJS NPM packages. Neither the use of Firefox Power Profiler or CO2.js is required for attendees and those unable to run them will still be able to get plenty of useful knowledge from this session.