After 5 years of data collection, our WARN-D machine learning competition to forecast depression onset is now LIVE! We hope many of you will participate—we have incredibly rich data. If you share a single thing of my lab this year, please make it this competition. eiko-fried.com/warn-d-machi...

Great talk by @hollycummins.com at Lindholmen developer day. As a data person and as someone who used to work with healthcare this resonated a lot: "don't make human tell computer things the computer already knows". It is so key but so often neglected in practice (&definitely a thief of joy) #databs

Thank you everyone for your suggestions! This is how I ended up plotting my LLM system output vs human annotation vs LLM-as-a-judge evals. Extra thanks to @libbyheeren.bsky.social for boosting my original question. #databs #dataviz

Any references for good visualizations of LLM evals? We have orig system output (yes/no), a human annotator's y/n and a number of different y/n evaluations. The aim is to see how good our original system is and how much of the issues do the diff LLM evals catch. How would you visualize this? #databs

#DataBS Conf 2025 preshow! We have two talks that we couldn't fit into the schedule but the speakers pre-recorded their talk for us to share before the main event next week! Both are really good and give me lots of excitement about what we'll see next week. ti.to/databsconf/d... <- free tix 🧵1/3

This week on Counting Stuff, the #dataBS conference attendee registration form is open! Also an intro to the 6 of 14 talks we have confirmed! Tickets are free/pay-what-you-want and there's technically a limit on attendees so grab yours today! www.counting-stuff.com/databs-conf-...

The weekend is here! Perfect time to submit your #dataBS talk. We want your lessons-learned stories from data work, everything from "my data pipeline: the unsung hero" to "how we got that AI system working" to "what we learned when it fell over." Details and sign-up: bit.ly/dataBSconf-cfs

Hey #dataBS! Got some stories about how data is messily done to accomplish real things? Have you learned a painful lesson about how not to do something? Then you've got something to share and should submit a talk to the Data Behind the Scenes Conference! (Form closes July 31) bit.ly/dataBSconf-cfs

The #dataBS (Data Behind the Scenes) Conference Call for speakers is out! We're gonna do this!! All online, single track, free to attend. Come talk about your messy experiences doing data stuff. At work, personal projects, whatever. A space to commiserate about nerdy things! bit.ly/dataBSconf-cfs

Academics, get thee some fuck off skills. Marketable skills that can help you fuck off from academia and get employed elsewhere. I see folx do multiple postdocs because they don't know what else they could do. Don't let that be you - make sure you have the ability to fuck off #academia #AcademicSky

I seem to re-learn this truth at regular intervals: if you're looking for an explanation of how something slightly niche works, find a master's thesis related to the topic. Students (at least in the nordics) are expected to explain the whole thing and not leave steps out (unlike academic articles).

I was studying at the uni and we had programming courses where you could get support from the TAs on the course's own IRC channel. I used Google reader to keep up with my favorite blogs. There was so little fluff or ads on websites, it was just raw html and original, stellar content 🥺

Ok, this is the kind of April Fools joke I love. Read the extracts below!

You will all stop everything you are doing and HAIL this CHONKY QUEEN

My first #tidytuesday #pydytuesday where I look at how different palmtree species are armoured. I challenged myself to work in plotnine and it's been a learning curve for sure! Code available here github.com/jtsuvile/tid...

Only dating advice I can give is to always check for implausible dates (like 1900-01-01 or anything >50 years into the future) as potential missing values in disguise. #dataBS #datasky

I'm trying to use a metric_set with accuracy together with cal_validate_logistic() from {probably}. I'm having a hard time figuring out how to set the parameter 'truth' (which typically is 'target'). @topepo.bsky.social do you know if there is a vignette showing how to do this somewhere? 🙏 #dataBS

I'm trying out a new thing in my collaboration with clinical researchers this time around: I created a Quarto document with a "results section" based on a dummy model and I'm using it to get buy-in for the metrics & plots before we actually fine-tune the model. Let's see if this works! #dataBS

Bluesky, we need this right now: see Umbrella, post Umbrella. www.youtube.com/watch?v=jPCJ...

This is something I see a lot - people who know very little about AI want to use it to do every thing. People who understand AI better are much more reserved. I do my best to disillusion folks in my organisation..

I no longer live in the country I grew up in. Every time I come back home I'm floored by how easy dealing with folks (customer service, contractors) is when I a) get to speak my native language and b) actually understand the rules that the society operates by. Just 🤯

"the methodology seems to work well enough that you can't rule it out immediately, *even if might be garbage for your particular task due to the specific intersection of quirks of your project.*" it's always the project specific quirks that get you.. #dataBS www.counting-stuff.com/labeling-thi...

Do you know a UK-based wheelchair user who works in a lab and would be willing to user-test our prototype of what is believed to be the first-ever lab coat adapted for wheelchair users? Please share far & wide! Interest form: forms.office.com/e/66FhcQjqRT More info: www.ucl.ac.uk/ucl-east/new...

I gave a talk earlier in the fall about AI in healthcare, talking about the hype and the challenges and where does that leave us. My talk focuses on a Nordic single-payer style healthcare but a lot of it generalises to different healthcare systems youtu.be/qqcYHWlCKkk?... #healthcare #dataBS

My colleagues are impressed by my shitposting game on the designated shitposting teams channel when really I just repost the best of #dataBS

...and just like in Harlow's monkey experiments, IRL interaction with conspecifics is much, much better than either of these

Playing around with the new Mall package mlverse.github.io/mall/ by @theotheredgar.bsky.social. Such a cool concept! But I cannot seem to figure out how to make my own data frame (CSV imported via polars) into the correct data type (which I suspect is MallFrame). Any tips? #dataBS

Loving the @nhsrcommunity.bsky.social #RPySOC conference and would warmly recommend it to all of us health/hospital #dataBS people. It's my second year following it online and the useful info / unit of time ratio is the highest I've found in this space nhsrcommunity.com/conference24...

"If I could also get added to the PR before you refactor the schema design that would be great"

Coming from academia (and #RStats) I'm used to building ny analyses with functions and pipelines. Reading Fluent Python has me thinking that maybe custom classes could be better for readability. What do you do #databs peeps - functions all the way or custom classes?

This is the fundamental paradox of AI: if it's actually helping you, there is no way to know when it is no longer helping you. Put another way: if you can supervise it effectively enough to catch its mistakes, you probably didn't need it in the first place.

"Google Translate can now render your college administrator’s opaque prose into plain (if terrifying) English." www.mcsweeneys.net/articles/fac...

types of data structures in computer science: FIFO: preserves order for queues LIFO: stack behavior, common in programming languages FAFO: the novel data structure that one guy on the team who seems to just show up when he feels like it invented which has a bunch of "DO NOT CHANGE" comments

my Likert Scale ⭐️⭐️⭐️⭐️⭐️: Lovert ⭐️⭐️⭐️⭐️: Likert ⭐️⭐️⭐️: Whatevert ⭐️⭐️: Dislikert ⭐️: Hatert

My One True Character Archetype is the cranky asshole who looks around and goes "fuck, either I handle this or someone who's gonna fuck it up worse than I will is gonna, god fucking dammit"

I told a colleague that logistic regression is AI and they got mad at me, so I made a chart. Find yourself. I am "Tinder is AI". (I posted this on that other site and a friend complained "y u no skeet this too!?!!?" so here you go)