I’m Lead Data Scientist at NHS Grampian and Senior Research Fellow at University of Aberdeen.
I run the Networked Data Lab, currently doing research on patient bottlenecks. We use big, messy, high-security health & social data. We make tools and analysis to help improve care.
I’m a huge nerd for open science and improving research culture.
| Website | https://jessbutler.github.io/ |
| https://twitter.com/JessButler284 | |
| Uni | https://www.abdn.ac.uk/iahs/people/profiles/jessicabutler |
Changes afoot for REF2028
Big differences in who submits and how much. And more weight given to _how_ we do research.
Best summary
https://www.timeshighereducation.com/news/ref-2028-all-you-need-know-about-submissions-and-assessment
Worthwhile meeting with the people who made the changes
https://www.eventbrite.co.uk/e/the-emerging-shape-of-ref-2028-tickets-656812392547
UKRI
https://www.ukri.org/news/early-decisions-made-for-ref-2028/
The mighty @dimitrablana has organised a great Data Science event
15 speakers from around Scotland will be talking about doing morally and technically sound work in data science
Thursday 1st June
join online or in person at OneTechHub
With breakfast, lunch, and lots of networking
https://www.abdn.ac.uk/achds/engagement/women-in-data-science-conference-414.php#panel418
The House of Commons has released a thoughtful, critical report on UK research quality
The recommendations to government are excellent, both practical and progressive
Highlights:
Make all research open access
Require data and code accompany all publications
Move peer-review and publication acceptance before data collection
Specifically fund replications
Remove “originality” from REF, instead require transparency
Impose a 3-year minimum term for all postdocs
https://publications.parliament.uk/pa/cm5803/cmselect/cmsctech/101/summary.html
Weirdly compelling challenge from the British Library
They give you a historical map, you have to pin it to the modern location. Can be anywhere in the world. No clues beyond the book title.
https://britishlibrary.georeferencer.com/maps/af6731d4-4032-41cc-8794-fad2719dadb8/georeference
Such a good presentation by Daniel Papasian and Todd Underwood.
A 10-year analysis of how machine learning pipelines break at Google
Punchline: it's rarely the machine learning. It's often a failure of testing the data and processing upstream.
100% applicable to non-google and non-machine-learning analysis pipelines
(h/t @vicki)
#NormConf is a tech industry conference focussing on stuff that matters in data and machine learning but doesn't get the spotlight.
They just released their 5-minute lightning talks
https://www.youtube.com/playlist?list=PLYXaKIsOZBstGUTXZXp2azDk8UJhpVVq3
I'm enjoying them. It's eye-opening how different industry is than academia on a lot of fronts.
🏒 Data goalie: keeping pucks out of the net
https://www.youtube.com/watch?v=MjPcMMeiR0o
🌳 Trying to convince academics to use git
https://www.youtube.com/watch?v=KgxKJjzh-Y4&list=PLYXaKIsOZBstGUTXZXp2azDk8UJhpVVq3&index=13
🦊 How to name files
https://www.youtube.com/watch?v=ES1LTlnpLMk&list=PLYXaKIsOZBstGUTXZXp2azDk8UJhpVVq3&index=6
Lots kids on the pickets today because their teachers are striking for fair pay too. Every school in the city is completely closed.
Me and the six-year-old made a "TAX BILLIONAIRES" sign. Her verict of the front line: really great snacks. Solidarity 😅
From UCUAberdeen:
