In May N8 CIR finally saw the final ReproHack of the Northern Tour take place. Originally due to take place on 14 March, but postponed due to the Coronavirus outbreak, the event was reconfigured for online delivery. This change saw the event become a one day mini-conference facilitated by the entire ReproHack team. Online delivery meant that all of the original participants could join the event, as well as up to 50 participants worldwide. This open registration saw participants from five continents join the event.
This brief overview will focus on the early part of the day and the extra talks but not the reproducibility hackathon as N8 CIR didn’t participate in those directly.
Introductions
The day began with an introduction from Anna Krystalli who outlined the structure of the event as well as giving an overview of the Blackboard collaboration platform that was used to deliver the event.
The platform worked incredibly well, enabling groups to communicate and work closely together on reproducibility but to come together quickly and effectively to discuss their findings, offer feedback or listen to the event’s speakers.
A quick warm-up saw participants outline the cartoon character with which they most closely identify. Answers were mainly from the Disney stable and included Baloo the Bear from Tale Spin, Chip or Dale from Rescue Rangers and Stitch from the feature film Lilo & Stitch.
With everyone introduced and aware of how supportive or destructive their colleague’s alter egos were it was time to hear from the day's first speaker
Daniel Nüst - Using Research Compendia
After Daniel's talk participants headed off to try their hand at reproducing some of the papers submitted to the event. On returning to the main room ahead of the lunch break there was time for some brief feedback of the initial reproduction attempts.
You can read more about Daniel's work and presentation here: https://n8cir.org.uk/events/event-resource/reprohack/nust/
Daniel Piqué - How I identified a missing data point in a paper with 8000 citations
Daniel Piqué who identified a missing data point in a paper that has been cited more than 8000 times. In the talk he spoke about his attempts to use R to reproduce the paper's only figure, how this led to spotting the missing data and also identified other errors in the work. You can read more about Daniel's research and presentation here: https://n8cir.org.uk/events/event-resource/reprohack/pique/
Paper Feedback
Following from such an interesting and engaging talk, the participants headed back to continue their reproduction efforts.
Participants gathered to give final feedback on the papers they attempted. Many people had difficulty in preparing the system environment of the paper, including libraries and software dependencies. Some papers had information or data behind journal paywalls and some were relatively easy to reproduce.
Sarah Gibson - Using Binder for Reproducible Research
The day's final talk was delivered by Sarah Gibson. She extolled the virtues of using Binder to reproduce the computational environment, which can help to mitigate many of the challenges faced by the participants on the day.
You can read more about Sarah's work and presentation here: https://n8cir.org.uk/events/event-resource/reprohack/gibson/
Conclusion
Ultimately the Remote ReproHack was an incredibly successful day with up to 50 participants joining the event at various stages. These participants spanned the globe so the event also offered great networking and learning opportunities. It is perhaps a glimpse into how future events may be organised and structured.
If you're interested in hosting your own ReproHack visit: