Prototyping tools to improve the understanding of science

Allon Scotton, Charles Andrews and Richard Hall, a team of developers from Thomson Reuters, were awarded the eLife challenge prize.

With their ‘ReLight’ hack, Allon, Charles and Richard aimed to help non-expert readers understand and learn more about biomedical research while reading research articles. They built a Google Chrome extension that identifies and highlights biomedical terminology in an eLife article, allowing the reader to access plain-language summary descriptions of the terms (via the Bing search API) as well as a list of related articles for those terms elsewhere on eLife. A key feature of their extension was the ability to guide the reader to additional information whilst remaining within the context of an article, a practical response to the often overwhelming deluge of information resulting from an unconstrained online search. The RedLight hack demonstrates the functionality that can be added to research articles by connecting together open APIs from different domains.

I applied [to Hack24] because I have wanted to take part in a hackathon and test my skills against others in a challenge format. I also wanted to see what I could come up with as part of a team that would make the world a better place. I think the fact that the challenge was about linking research together resonated with our desire to use the Thomson Reuters and the Microsoft APIs. Hack24 gave me increased confidence in my skills as a developer and it was great to see what can be created in just 24 hours.

– Allon Scotton, Senior Software Engineer, Thomson Reuters.

In their hack, Allon, Charles and Richard used the eLife search API, Microsoft Cognitive Services, Thomson Reuter’s Open Calais API, and the Bing search API. The back end was written in Python and Java; the front-end in JavaScript. Moving forward, the team feel they could improve the tool’s accuracy as well as expand it to work on other online content and for a wider range of terminology. The code for the hack is online at https://github.com/a-pns/Hack24-Last_Minute.

Another notable hack developed for the eLife challenge aimed at connecting people reading about science in the news to the underlying research, to help build an informed understanding of the latest discoveries. Joshua Shammay, Luke Snopkiewicz and Richard Paul created a Mozilla Firefox extension for retrieving research articles from PubMed directly from the news webpage, thereby connecting a wider audience directly with the research literature. An interesting feature of this hack was the addition of community curation of relevant articles via a simple upvoting mechanism: the team explored developing this into a mechanism to indicate which articles support the statements in the news article.

“CuneiPhrase (a shortening of Cuneiform Phrases) is a system for public knowledge engagement, intelligently bridging news and research through AI and the wisdom of the crowd. A user visiting a site can upvote and downvote on related journal articles, suggested by AI or found via search, and see the relations that have been ranked highest by the community. This hack addresses the difficulty of sourcing research that is related to or underlying the things we read on the internet — more relevant today than ever.”

– Joshua Shammay, VP of Web Development, Decorilla.

The team used Microsoft Cognitive Services to extract intelligent keywords from online news articles and the PubMed search API to find articles with these keywords. The code for the extension is available at https://github.com/soundcake/CuneiPhraseBackend (back end) and https://github.com/soundcake/CuneiPhrase (front end).

While many other ideas were not ultimately developed into hacks for this event, we received suggestions from several participants with prior experience of research and medicine who were interested in the problem domain we laid out with our challenge. These included formalising a protocol for peer review inspired by internet protocols; understanding the credibility of an article based on the ongoing validity of the research it builds on;and creating a Git-like forking mechanism that allows the community to see how data is being reused more quickly than by formal citation, as well as facilitating communication between data authors and data reusers early in the process.

There was a great degree of creativity and skill demonstrated in many of the hacks submitted. All the winners from this year’s Hack24 are listed at https://www.hack24.co.uk/2018-winners/.

Finally, teamwork at Hack24 was helped in no small way by the Slack bot created especially for this event by Tech Nottingham community members, led by David Wood (@codesleuth). Hackbot is an open source bot that allows participants to create new teams, with name and motto, and add and remove participants from these teams.

“We have some exciting plans for Hackbot which will bring it squarely into 2018 by using natural language processing and offering it as software-as-a-service to other hack events. We plan to make use of Hackbot to smooth out the judging process and give teams the tools they need to submit their hacks to any challenge they wish, as well as support these features with tools for hack event organisers.” - David Wood

The team at Tech Nottingham encourage others to adopt Hackbot for similar events and welcome contributions to extend its capabilities. You can get in touch with them on GitHub or the Tech Nottingham Slack channel #hack24-hackbot-coders at technottingham.slack.com.

Scripts: https://github.com/TechNottingham/hubot-hackbot

Back-end API: https://github.com/TechNottingham/Hack24-API/

Shell: https://github.com/TechNottingham/Hackbot

How to use it: https://www.hack24.co.uk/how-to-use-hackbot

Do you have an idea or innovation to share? Send a short outline for a Labs blogpost to innovation@elifesciences.org.

For the latest in innovation, eLife Labs and new open-source tools, sign up for our technology and innovation newsletter. You can also follow @eLifeInnovation on Twitter.