Lab Meeting on 15 August 2019

We had so many fascinating conversations this week that we didn’t get to spend as much time as we’d planned discussing how to make lab meetings effective and useful for all.

It seemed like interrupting team members sharing their expertise to talk about how to build a space where team members shared their expertise didn’t make much sense at all 😉

Check out all the links below! We hope they’re useful to you ✨

Celebrations and cool things to share

Kirstie had a wonderful time at Neurohackademy and shared the YouTube playlist that has all the talks from all four years of the summer school. She was also delighted to celebrate Gamechangers for Diversity in STEM being featured in the Turing annual report! (Check out pages 35 to 38.)

Maxine is enjoying writing up her thesis and a short break to come and say hello in person 👋

Sarah is leading a “Learn Azure in a Month of Lunches” skills development group with some collaborators at the Turing. Should be fun, and yummy 😋

Louise lead a discussion of how useful git blame can be, and how terrible the name is for making people feel welcome and supported! git blame lets you find out in which commit an edit to a particular function was made and she’s found it really useful for figuring out the context of a change when the file itself was too big to go through the commit history manually. There are some really interesting conversations about potentially renaming the function, for example in the GitLab UI.

Georgia is really excited about the Autistica/Turing Citizen Science project building up. She’s organising focus groups for September and working with collaborators to get the demo developed into a minimum viable product so we can run user testing interviews 🚀🌟

Ang shared beautiful visualization tool in Python: Altair. Have a play yourself on Binder.

Malvika is hosting a statistics workshop followed by a Machine Learning in R course later this year and shared a lovely Introduction to Machine Learning resource with tutorials, exercises and videos by Bernd Bishl and his team.

Alex shared Daft: an amazing package to beautifully render probabilistic graphical models.

Questions we’re thinking about

Louise asked the group their opinions on tutorials to learn git. The Software Carpentry lesson is very famous but some instructors don’t find it very useful for new learners. Kirstie recommended Elizabeth’s tutorial from Neurohackademy and the recording of her talk, and the Friendly GitHub Intro workshop that she’s run multiple times.

Sarah is looking for anyone with expertise in kube-lego and Let’s Encrypt! (No one was able to help her 😭)

We did have some suggestions around ways to manage her TODO lists though.

Maxine and the One Health Tech community are looking for international terms to monitor ethnic diversity at their events. The best solution we found was to use just the top level from the UK groupings but we did also discuss tools that can summarise free text for the specific purpose that is needed. For example the GenderCoder tool lets people enter their gender in free text so they can self identify, and also facilitates grouping free text answers that are all versions of the same gender (eg: M, male, a man, Male are all grouped together.)

This led really well onto a discussion that Georgia raised about monitoring diversity for her project. Some suggestions were to have continuous monitoring and looking at diversity in leadership. In general the take home message was that recording nothing makes reflections almost impossible in the future!

Yini asked for the group’s opinion on two journals (NeuroImage and Human Brain Mapping) going open access. Kirstie had a strong opinion that these are business model changes from the publishers and the open access fees do not make the journals all that more “open” or “accessible”. In fact, in writing up this blog she found a recent blog about “Why open access will boost publisher profits” which makes that point really clearly.

Alex asked for tips and tricks for reviewing the literature when starting out on a new project. We recommended this 10 Simple Rules paper by Marco Pautasso in 2013 (doi: 10.1371/journal.pcbi.1003149)

Rethinking lab meeting

Although the discussion points above filled a lot of lab meeting [😻 – KW] we still discussed Kirstie’s proposed guidelines for lab meeting.

They explain the process and motivation for bringing groups of people who all work very closely with Kirstie together. We’ll finish this post with a shout out to all the lab meeting blogs so far and the really fantastic conversations that they have inspired, and a paragraph about the core reason why Kirstie is so passionate about having everyone participate every week.

Trust and the power of decentralised networks

(Copied from the lab meeting guidance and used under CC-BY license.)

There is a great risk that the projects in the group are siloed and that Kirstie becomes a bottle neck for communication. This is a centralised network (A in the diagram below). In contrast, a decentralised network (B) shares information more effectively. Small world networks in particular are efficient and resilient.

Centralised and decentralised networks
Decentralization diagram uploaded to Wikimedia Commons by Adam Aladdin. Reused under a CC-BY license.

I would like this lab to be as decentralised as possible. I would like lab members to find the answers to their questions quickly and receive help and support from all other lab members in all dimensions.

One of the core goals of all Whitaker Lab is to improve collaboration (in academia and in the world!) And a core requirement of effective collaboration is trust. In fact, there was an article in the Harvard Business Review in 2011 that named trust as the “The One Thing That Makes Collaboration Work”.

Building trust is a difficult and multi-dimensional, but by being vulnerable with each other is one core component. In her book Emergent Strategy (which I love love love), and in this blog post from 2013, adrienne maree brown describes the following principles:

  • lao tzu says “if you don’t trust the people, they become untrustworthy.” the first principle is a positive flip of this statement – if you trust the people, they become trustworthy. trust is a seed that grows with attention and space. the facilitator can be a gardener, or the sun, the water.
  • there is a conversation in the room that wants and needs to be had. don’t force it, don’t deny it. let it come forth.
  • the connection between the individuals is what makes the whole group/community strong.

(Note that brown self determines what she capitalises, which reminds me of - and is possibly inspired by - black feminist writer bell hooks who did not capitalise her name in order “to focus on the substance of her work, not who [she] is”.)

If we want information to flow around a decentralised network, then we need to trust each other. I believe that building connections between individual people are the magic that will lead to the most exciting and innovative data science.