Ben Harrap

Ben Harrap

@bharrap.bsky.social

doer of data things at Yardhura Walani president of statsoc.org.au I do #rstats, surveys, biostatistics, #IDSov on Ngunnawal and Ngambri lands not Indigenous, non-binary (they/them) benharrap.com

2.4K followers826 following1.4K posts

Pinned

Ben Harrap
@bharrap.bsky.social

Re-sharing my starter pack so I can update my pinned post If you're a statistician, data scientist, or someone doing statsy or data sciencey things in Australia or Aotearoa (New Zealand) and want to be on the list, just let me know! go.bsky.app/3wERdAo

Ben Harrap
@bharrap.bsky.social

7pm: kiddo in bed, lemme just open my laptop and reply to an email real quick 11pm: whoops I just finished a manuscript 😵‍💫

Ben Harrap
@bharrap.bsky.social

Hadley's post clarified for me the role the language we use plays in describing AI "collaborate" and "expertise" are the kind of terms I'd use for a person, not a tool (you wouldn't collaborate with a hammer) AI might be unlike any other tool we've created, but it's still just a tool

Hadley Wickham@hadley.nz

I made an AI video about genzplyr and I lowkey love it. It's also 100% slop. In this post I dig into what makes something slop, when it's ok to share it, and how to avoid it in work that matters. Read more at: substack.hadley.nz/p/what-is-slop

substack.hadley.nzWhat is slop?I made an AI video about genzplyr and I lowkey love it. It's also 100% slop. In this post I dig into what makes something slop, when it's ok to share it, and how to avoid it in work that matters.
Ben Harrap
@bharrap.bsky.social

The most underrated thing about teaching Intro to R is that you get to choose how cohorts of #rstats users will go on to pronounce package names dplyr? dplyr? Or is it dplyr? They don't know, but they don't know that you don't know either!

Ben Harrap
@bharrap.bsky.social

"first install the tiddyverse" "when working with factors, I'd recommend fokats" "dipler is full of useful functions" "characters are best manipulated using strin-grrrr"

Ben Harrap reposted

Charlie Gao
@shikokuchuo.net

I am so looking forward to giving this webinar for the Statistics Society of Australia. I'll be offering a glimpse into what I'm working on right now for the next generation of parallelism. The event is still open for you to sign up - 8am BST / 9am CEST Thursday 22nd Nov - link below. #Rstats

Fonti Kar@fontikar.bsky.social

Level up R workflows with @shikokuchuo.net 🚀 Charlie will give us a practical tour of the open source stack he created to improve parallel computing 📦 Join our Stats Computing + Data Vis Webinar and learn from the best💡 🗓️ THU 22 Oct ⌚️18:00 AEDT | 07:00 UTC 📌 Register statsoc.org.au/event-6817896

Ben Harrap
@bharrap.bsky.social

What the enshittification is this 'AI Assistant' update to Adobe Acrobat I just want to fill in a form, but instead I'm harassed by - A top banner telling me to 'save time by reading an AI summary' - A sidebar taking up 1/3 of the window - A bottom banner with the option to "hide for this session"

Ben Harrap
@bharrap.bsky.social

With the new `password = ` argument I can now brute-force password-protected .xlsx files all within the tidyverse muahahaha #rstats

Jenny Bryan@jennybryan.bsky.social

🖇️ readxl 1.6.0 is released 🖇️ It includes fixes for several CVEs and can read password-protected xlsx files 🗝️ #rstats readxl.tidyverse.org/news/index.h...

readxl.tidyverse.orgChangelog

Ben Harrap reposted

Fonti Kar
@fontikar.bsky.social

The time has come! Our annual invited webinar at SCV will be featuring wonderful @shikokuchuo.net! Join the community and come along to learn about the latest developments in parallel computing in R 🗓️ THU 22 Oct ⌚️18:00 AEDT | 07:00 UTC 📌 Register statsoc.org.au/event-6817896

Fonti Kar@fontikar.bsky.social

Level up R workflows with @shikokuchuo.net 🚀 Charlie will give us a practical tour of the open source stack he created to improve parallel computing 📦 Join our Stats Computing + Data Vis Webinar and learn from the best💡 🗓️ THU 22 Oct ⌚️18:00 AEDT | 07:00 UTC 📌 Register statsoc.org.au/event-6817896

Ben Harrap
@bharrap.bsky.social

The suggestions weren't exactly what I had in mind, so I put together something to illustrate. Clips in the thread showing some of the features At the time I thought it was a good/useful idea, but is it actually? Disclaimer: All DAGs/studies are completely fabricated! It's just to demonstrate

Ben Harrap@bharrap.bsky.social

Day 1 of #AEA2026 off to a great start Talk by Brett Dyer (Griffith Uni) got me thinking - is there a repository that pools published DAGs to show collectively what is thought to be related to a given outcome? Like a Cochrane review for DAGs @pwgtennant.bsky.social @epiellie.bsky.social #episky

View full thread
Ben Harrap
@bharrap.bsky.social

Final disclaimer - this is a proof of concept, put together using Claude over an afternoon. Several features don't work (e.g. 'uploading' a DAG only stores it in your browser session, I don't keep them anywhere)

Ben Harrap
@bharrap.bsky.social

The suggestions weren't exactly what I had in mind, so I put together something to illustrate. Clips in the thread showing some of the features At the time I thought it was a good/useful idea, but is it actually? Disclaimer: All DAGs/studies are completely fabricated! It's just to demonstrate

Ben Harrap@bharrap.bsky.social

Day 1 of #AEA2026 off to a great start Talk by Brett Dyer (Griffith Uni) got me thinking - is there a repository that pools published DAGs to show collectively what is thought to be related to a given outcome? Like a Cochrane review for DAGs @pwgtennant.bsky.social @epiellie.bsky.social #episky

View full thread
Ben Harrap
@bharrap.bsky.social

It's also possible to show interact on a study-level, such as omitting it from the compiled/strict DAGs, or showing only that study's DAG.

Ben Harrap
@bharrap.bsky.social

The suggestions weren't exactly what I had in mind, so I put together something to illustrate. Clips in the thread showing some of the features At the time I thought it was a good/useful idea, but is it actually? Disclaimer: All DAGs/studies are completely fabricated! It's just to demonstrate

Ben Harrap@bharrap.bsky.social

Day 1 of #AEA2026 off to a great start Talk by Brett Dyer (Griffith Uni) got me thinking - is there a repository that pools published DAGs to show collectively what is thought to be related to a given outcome? Like a Cochrane review for DAGs @pwgtennant.bsky.social @epiellie.bsky.social #episky

View full thread
Ben Harrap
@bharrap.bsky.social

Another idea was to filter in/out nodes and edges based on some criteria Number of times a variable appears across DAGs could indicate agreement across studies Publication date could show evolution of understanding over time

Ben Harrap
@bharrap.bsky.social

The compiled DAG view just combines them all, meaning cyclical and bidirectional relationships are shown. The strict view is meant to comply with DAG principles. Can also toggle to add weight to relationships that are more commonly specified

Ben Harrap
@bharrap.bsky.social

The suggestions weren't exactly what I had in mind, so I put together something to illustrate. Clips in the thread showing some of the features At the time I thought it was a good/useful idea, but is it actually? Disclaimer: All DAGs/studies are completely fabricated! It's just to demonstrate

Ben Harrap@bharrap.bsky.social

Day 1 of #AEA2026 off to a great start Talk by Brett Dyer (Griffith Uni) got me thinking - is there a repository that pools published DAGs to show collectively what is thought to be related to a given outcome? Like a Cochrane review for DAGs @pwgtennant.bsky.social @epiellie.bsky.social #episky

Ben Harrap
@bharrap.bsky.social

What I'd imagined was a way to compile all DAGs relating to the same outcome in a single mega-DAG. The publications that it's drawn on are shown on the right, and clicking on a variable highlights the studies that include it #episky #academicsky #econsky

Ben Harrap reposted

Adrian Barnett
@aidybarnett.bsky.social

A major Australian funder will now use a conditional lottery to award funding. I think this is a great policy change and lotteries are likely to become more common as AI-driven application numbers grow theconversation.com/some-austral...

theconversation.comSome Australian research funding is now a lottery. And that’s actually good newsLeaving some funding decisions up to luck can save everybody time and effort – and help understand what the funding actually does.
Ben Harrap
@bharrap.bsky.social

Nick's a great educator - if you're looking to get started with {targets}, this is the workshop for you!

Nick Tierney@njtierney.bsky.social

Hey folks! Stoked to be giving a 3 hour workshop "A Gentle Introduction to Targets" at WOMBAT in Melbourne, November 30. Sign up at wombat2026.numbat.space/program/tuto... Early bird registration prices ends Sunday, 18th October! See the full program wombat2026.numbat.space #rstats #targets

Ben Harrap
@bharrap.bsky.social

Between conference sessions and social events, you're basically doing 12hr days All the while your inbox piles up, so when the conference ends you're stuck playing catch up I like conferences but I do not like this part >:(

Ben Harrap
@bharrap.bsky.social

Good analysis isn't enough to make change - we need to. Number 1 tip - connect health data with the economics of it when producing policy advice

Ben Harrap
@bharrap.bsky.social

While the screen was comically small to be presenting from, my #AEA2026 talk went pretty well! Choosing to have few words, big font, and big pictures paid off

Ben Harrap reposted

Fonti Kar
@fontikar.bsky.social

💭 Have you built software & wondered how to share it with others? 💭 Feel overwhelmed by writing a publication to show-off your work? 💡Join our Publishing Your Software tutorial at #WOMBAT2026 🗓️ 30 Nov 1330 📌 Register lnkd.in/gHj2HQwe @visnut.bsky.social and I would love to see you there! 😁