The main theme was all about analytics and metrics. Participants were encouraged to bring and work on their own ideas or join forces with other teams. We also provided them with some ideas on what to work on, e.g. using metrics to create trend or gap analysis (“What's hot and what's not in science”), create new or calculate alternative metrics to challenge the Journals Impact Factor or explore how analytics and metrics can help with solving Grand Challenges such as Climate Change, Sustainability or Global Health.
This time, we used various communication channels to attract potential participants. We created dedicated e-mail invitations and mailing campaigns, and also announced the event on our social media channels and to our Institutional Sales colleagues. Before the event, 51 people signed up on our Eventbrite registration page. For the actual hack day, more than 40 participants showed up, with a good balance of researchers, developers and publishing professionals that worked out very well.
This time with more time
In comparison to previous events we decided to try out a new format for this Hack Day. Instead of hosting it for only 1 day, we extended the event to 1.5 days. We started off on Thursday afternoon and evening with introductory sessions, brainstorming and socializing. Friday was reserved for the actual hacking. This concept gave our attendees more time to acquaint themselves with the topics, the different working groups, and the available data sets. Almost all participants joined us for the social dinner on Thursday evening at the Heidelbeere next to our office where they continued to network and
The course of the days
Henning Schoenenberger kicked off the event by introducing the strategy behind the Springer Nature Hack Day series and elaborating on the aims of this event. Afterwards, our external guest speaker Francesco Osborne from the Knowledge Media institute (KMi) at the Open University in the UK gave a talk about the newly released “Computer Science Ontology – A large-scale taxonomy of research topics” and how it helps with discovering trends and doing analytics in Computer Science.
Before the participants joined in a lively discussion around potential hacking projects, they had the chance to pitch their ideas on the stage. Jan Steinkraus and Chris Bendall from Springer Nature took the opportunity to pitch their use cases for Springer Medizin and Corporate Markets respectively, and some participants from the audience followed their example and presented their ideas. The purpose of these pitches was to give attendees who didn’t know yet what to work on an incentive to join their teams.
This brainstorming session worked out well. On Friday morning, we didn’t even need an official kick off - the participants eagerly started directly with their projects, only interrupted by the need for caffeine or carbs.
Participants could choose from an astonishing variety of different data sets we made available to them:
- Springer Nature SciGraph: Over 1 billion triples (articles, abstracts, books, grants, conferences, etc.) available for download as Linked Open Data and via API.
- Springer Meta(data) APIs: metadata feeds for almost 12 million online documents (e.g. journal articles, book chapters, protocols).
- Springer full text API: access to full-text content for Springer articles and chapters
- Bookmetrix API: Downloads, citations, readers on Mendeley, mentions, reviews for 465 thousand books and for 4.6 million chapters
- UNSILO APIs: ‘Related’ returned top matching articles, ‘Metadata’ contains all the concepts that have been extracted from a given article
- DIMENSIONS API: Links publications, grants, patents and clinical trials in a consistent network of over 128 million documents, connected by approximately 4 billion links.
It’s worth mentioning that data traffic during this hack day was considerably higher than during the two previous events. Especially the Springer Meta(data) and Springer full text APIs were so popular with our participants that we had to increase the quota twice!
Result presentations and voting
After a productive second day everybody was very eager to find out what had been developed in the past 1.5 days. Every team presented their accomplishments on the stage, and the whole group had the chance to vote for their favorite projects. We decided to try something new here as well. Each participant had three votes, one per project / use case that had been presented in the following categories:
- Most innovative idea
- Best presentation
- Best real life applicability
The overall winner in all three disciplines was the “Disease Dashboard” from Springer Medizin, created by a large and very diverse group of participants. They focused on various analytics tiles that practicing doctors would need for a holistic look at a certain disease.
Surprisingly in second place was the “Semantic Entity Marker”, created and developed by our colleagues Michael Wolowyk, Silvia Pina and Ruba Kharrubi. They received a great score for the most innovative project: a Chrome browser extension that enables analytics by enriching articles with extra useful information from DBpedia, DIMENSIONS, Springer Nature SciGraph and UNSILO.
The “Hot Topics Analysis” corporate use case finished up in the third place, with great insights discovering key topics that are being focused on by researchers in the Tech industry. The team combined the Computer Science Ontology terms with download and denial data from Springer Nature. More information can be found in this great blog post by Angelo Antonio Salatino. There of course were other great projects as well. If you're interested you can find more information in the slide deck we'll make available once we have updated the official event site on springernature.com.
Post hack day
While a short highlights video will follow shortly we already have the first results from the survey we had sent out to participants. We are eager to find out what they liked and what could be improved in future events. 17 attendees participated which represents almost half of all attendees. They provided us with valuable feedback and input for future events. Here are some interesting facts we were able to extract:
- Event length: 100% said that 1.5 days represented the perfect amount of time
- General rating of this Hack Day: 52,9% rated it “Very good”, 47,1% rated it “Good”
- Data quality, data coverage and data APIs were rated as very good and adequate
- Relevance: 82,4% rated the relevance of the Hack Day either “highly relevant” or “mostly relevant”
- Recommendation: 100% would recommend the Hack Day to a colleague
- Interest in future events: 76,5% would be interested in attending another Springer Nature Hack; 23,5% would consider to do so
Many thanks to all
I would like to extend my biggest gratitude to all participants for turning this into such a productive event. We’re already looking forward to bringing this event series to another venue. And, of course, an overwhelmingly big thanks and congratulations to all colleagues involved in the planning committee for the great work they did to ensure we had lots of people in the room, everything went smoothly and the event as a whole was a great success.