How Renku is using rich-text editing to take on the data reproducibility crisis.
When two branches of the Federal Institute of Technology in Zurich and Lausanne got together to create the Swiss Data Science Center (SDSC) and its Renku platform, the aim was to build confidence in data and results while ensuring that credit goes to the scientists who put in the hard work compiling them. With CKEditor 5, on which Renku was built, the SDSC finds itself as the creators of a powerful tool that can help scientists track and organize their data and analyses, no matter how complex they get.
Challenge: Compiling clear and credible data
When we spoke with Sekhar Ramakrishnan, senior developer for the Renku platform, he told us about the primary aims behind its creation. First, simplifying communication and collaboration within data science projects. Communicating effectively means knowing how to style and format text for the task at hand while also making it easy to do. SDSC originally considered Markdown for its text editing, but they found that the target audience for Renku was very uncomfortable with it. “We had users who were comfortable with Markdown, but most of them, about 70% actually, wanted a WYSIWYG editing experience,” says Ramakrishnan.
All of which aids Renku with its second aim: increasing the visibility and viability of data science. Having a rich-text editor with easy-to-use styling features such as tables and lists encourages precise and expressive communication about data, analysis techniques, and the final results and findings. Accurate reports of data are the key to realizing greater transparency in the process and increasing confidence in results. Unfortunately, there is an ongoing data reproducibility crisis and scientists are finding it difficult to reproduce previously published results. Causes for this include factors like the intense pressure to be prolific in publishing findings, as well as the pressure to please multiple stakeholders in a project. Therefore, clear conversation and transparency about original research are paramount for Renku, and CKEditor’s technology helps enable it.
Solution: A customizable rich-text editor that works together with Markdown, community support, and easy implementation
— it lets researchers share raw data as well as code to make multiple analyses possible. You may already know about GitHub Writer, CKEditor’s add-on for GitHub that lets users switch between Markdown and WYSIWYG. Because CKEditor 5 is easy to customize, the Renku team created a component that gives its users the ability to switch between WYSIWYG and Markdown editing as they wish. This means that in contributing, whether addressing data inconsistencies, explaining the choice of machine-learning algorithm, or debating the meaning of findings, researchers and analysts can work in a way that is intuitive to them. The rich-text editing features that CKEditor 5 provides, such as styles, tables, and lists, gives users the power to design text as they see fit for coding repositories.
SDSC computer scientist Virginia Friedrich was also impressed by CKEditor 5’s capabilities, being familiar with our rich-text editors from previous projects she was involved in. Friedrich decided to customize Renku’s editor to include Markdown and give them a pre-configured base to start from. “Some of our users are advanced users, and prefer to write in Markdown. We had to provide a way for them to switch between styled text and raw Markdown depending on their preferences,” she says, adding that being able to pull CKEditor code and customise the editor to fit Renku’s needs is a big plus. Whether reproducing and verifying previous results, or producing variations with existing datasets, one thing is clear: ensuring that Renku makes it easy for researchers to describe what they have done, and for others to continue the work if necessary to get it right, is vital for the advancement of science.
Friedrich additionally praised the ease of CKEditor’s integration process and the help our support team gave SDSC. “CKEditor is a great solution for us. It saved us a lot of development time, works great, and I think it will keep evolving to fulfil our current and future needs. I can see on GitHub that there is a very active team behind this product.” Ramakrishnan, also wanting an easy integration, chimed in. "I didn't want to integrate something that would become orphaned. I wanted our editor to be a live project." As CKEditor is always updating and evolving, this is not an issue for the Swiss Data Science Center.
We had users who were comfortable with Markdown, but most of them, about 70% actually, wanted a WYSIWYG editing experience.
Result: A business model with untold potential
The ability to reproduce scientific results has significant ramifications for the business world, especially as data science becomes a major factor in corporate decision-making. The SDSC has made Renku freely available for any business or research institution that wants to use it.
“We see a lot of potential for its adoption in industry,” Ramakrishnan says. “Data science in business really involves a lot of collaboration between multiple parties — data scientists, programmers, those with business knowledge, and also project managers to see that metrics for success are reached.”
Granted, Ramakrishnan admits that a lot of CKEditor features are still being integrated into Renku, but our conversation with him displayed his vision for the future in full view. Our WYSIWYG technology opens up possibilities to get data right with its seamless integration and its intuitive user interface, making clear layouts of data quick and painless. As Renku goes on to be a major platform for science, academia, and business alike, CKEditor 5 will remain the bedrock of this exciting initiative.
“CKEditor is a great solution for us. It saved us a lot of development time, works great, and I think it will keep evolving to fulfil our current and future needs.”
Swiss Data Science Center