David Asboth
@davidasboth.com
Data generalist, educator, author of The Well-Grounded Data Analyst (Manning, 2025). Co-host of the Half Stack Data Science podcast.
Pinned
David Asboth
@davidasboth.com
· Oct 29
Intro time! I'm David. I call myself a "data generalist" because I haven't found a label that fits.
I've been a software dev, data scientist, and my real passion is education so now I do Python and data training.
I'm currently most interested in skills that data people *actually* need to know.
I've been a software dev, data scientist, and my real passion is education so now I do Python and data training.
I'm currently most interested in skills that data people *actually* need to know.
A lot of the AI culture wars seem to stem from:
1) calling everything under the sun "AI", lumping ChatGPT in with random forests
2) not making a distinction between AI the technology and the AI that's the companies stealing copyright and shoving chatbots where they don't belong
#databs #datasky
1) calling everything under the sun "AI", lumping ChatGPT in with random forests
2) not making a distinction between AI the technology and the AI that's the companies stealing copyright and shoving chatbots where they don't belong
#databs #datasky
Is this platform still massively against AI or has it moved more towards acceptance?
November 26, 2025 at 8:43 AM
I can't explain it, but one of my comfort viewings is panel shows where Kevin Bridges is a guest and none of the other guests understand a word he's saying.
A fine example: www.youtube.com/watch?v=9cK_...
A fine example: www.youtube.com/watch?v=9cK_...
November 26, 2025 at 8:33 AM
I can't explain it, but one of my comfort viewings is panel shows where Kevin Bridges is a guest and none of the other guests understand a word he's saying.
A fine example: www.youtube.com/watch?v=9cK_...
A fine example: www.youtube.com/watch?v=9cK_...
Important thread. I think I only really rediscovered the power of play after becoming a parent, which is a shame because I should never have forgotten! Now, as an educator, I lean towards the playful and fun as much as possible. Life's too short not to have as much fun as you can.
Of course I would say this, but we need to talk about the moral & ideological case for play & leisure more than ever. So much political discourse - especially from the super-rich - assumes we ought to construct society around forcing citizens to spend most of their one, unrepeatable life working.
November 24, 2025 at 11:44 AM
Important thread. I think I only really rediscovered the power of play after becoming a parent, which is a shame because I should never have forgotten! Now, as an educator, I lean towards the playful and fun as much as possible. Life's too short not to have as much fun as you can.
Jazz aficionados of Bluesky! Can you recommend a jazz album for me and my friend's "album club" (like a book club but for jazz albums)? It's my turn to pick and I'd like some inspiration!
November 22, 2025 at 12:54 PM
Jazz aficionados of Bluesky! Can you recommend a jazz album for me and my friend's "album club" (like a book club but for jazz albums)? It's my turn to pick and I'd like some inspiration!
This is an incredible fact. It might even replace my current favourite fact, which is that Gary Numan is older than Gary Oldman (also an excellent fact).
8675309 is prime, and so is 8675311, so if you ever need a middlin'-large pair of adjacent primes to test your cryptographic suite, all you need is a 1980s earworm and a +2 and you're all set.
Man, everything is so bleak, anyone got a fun fact or little bit of trivia they want to share
November 21, 2025 at 6:58 AM
This is an incredible fact. It might even replace my current favourite fact, which is that Gary Numan is older than Gary Oldman (also an excellent fact).
Honestly it's a breath of fresh air seeing students pissed off by lecturers using AI-generated slides especially when there's a double standard whereby students can't use the same tools.
www.theguardian.com/education/20...
www.theguardian.com/education/20...
‘We could have asked ChatGPT’: students fight back over course taught by AI
Staffordshire students say signs material was AI-generated included suspicious file names and rogue voiceover accent
www.theguardian.com
November 20, 2025 at 2:14 PM
Honestly it's a breath of fresh air seeing students pissed off by lecturers using AI-generated slides especially when there's a double standard whereby students can't use the same tools.
www.theguardian.com/education/20...
www.theguardian.com/education/20...
If I get another email from Dropbox telling me my ❗storage ❗is ❗nearly ❗full ❗when I have AN ENTIRE GIGABYTE FREE, I'm gonna start flipping some tables. (I realise they probably calculate it on percentages, but some common sense wouldn't go amiss)
November 18, 2025 at 4:40 PM
If I get another email from Dropbox telling me my ❗storage ❗is ❗nearly ❗full ❗when I have AN ENTIRE GIGABYTE FREE, I'm gonna start flipping some tables. (I realise they probably calculate it on percentages, but some common sense wouldn't go amiss)
Vindication that entity resolution is still a Hard Problem in data science. We're still throwing all our latest tools at it!
#databs
#databs
Entity resolution in real-world datasets remains a persistent challenge.
A #preprint introduces a multi-agent Retrieval-Augmented Generation (RAG) framework that decomposes household entity resolution into coordinated and specialized agents.
📌 Full paper: buff.ly/xVpcKaV
#DataScience #dataBS
A #preprint introduces a multi-agent Retrieval-Augmented Generation (RAG) framework that decomposes household entity resolution into coordinated and specialized agents.
📌 Full paper: buff.ly/xVpcKaV
#DataScience #dataBS
Multi-Agent RAG Framework for Entity Resolution: Advancing Beyond Single-LLM Approaches with Specialized Agent Coordination
buff.ly
November 18, 2025 at 8:42 AM
Vindication that entity resolution is still a Hard Problem in data science. We're still throwing all our latest tools at it!
#databs
#databs
Yes yes yes!
Getting nervous for the talk I'm about to give at a workshop about "using AI to drive impact" which features slides such as these.
November 7, 2025 at 8:26 AM
Yes yes yes!
"You don't have to burn books, do you, if the world starts to fill up with non-readers, non-learners, non-knowers?"
Fahrenheit 451, afterword (Ray Bradbury, 1953)
#AI
Fahrenheit 451, afterword (Ray Bradbury, 1953)
#AI
November 5, 2025 at 9:19 AM
"You don't have to burn books, do you, if the world starts to fill up with non-readers, non-learners, non-knowers?"
Fahrenheit 451, afterword (Ray Bradbury, 1953)
#AI
Fahrenheit 451, afterword (Ray Bradbury, 1953)
#AI
A movie you've seen more than seven times with a gif.
November 4, 2025 at 7:09 AM
A movie you've seen more than seven times with a gif.
The Guardian's "how deprived is your area" interactive page is pretty neat, but the best part is this note:
"A technical issue in an earlier version meant localities containing a comma produced incorrect results."
No avoiding the bad data!
Here it is: www.theguardian.com/society/ng-i...
#databs
"A technical issue in an earlier version meant localities containing a comma produced incorrect results."
No avoiding the bad data!
Here it is: www.theguardian.com/society/ng-i...
#databs
How deprived is your area? Look up your postcode as new data for England released
New data ranks every area of England against a set of metrics for deprivation. Find out where yours figures in the statistics
www.theguardian.com
October 30, 2025 at 7:15 PM
The Guardian's "how deprived is your area" interactive page is pretty neat, but the best part is this note:
"A technical issue in an earlier version meant localities containing a comma produced incorrect results."
No avoiding the bad data!
Here it is: www.theguardian.com/society/ng-i...
#databs
"A technical issue in an earlier version meant localities containing a comma produced incorrect results."
No avoiding the bad data!
Here it is: www.theguardian.com/society/ng-i...
#databs
Another fun data problem with my bot.
The Wikipedia data has both Serbia & Montenegro AND Yugoslavia as opponents of the same game, so my SPARQL query picked them up as two separate games.
Yugoslavia itself didn't even exist anymore in '97 but apparently the football team still used the name!
The Wikipedia data has both Serbia & Montenegro AND Yugoslavia as opponents of the same game, so my SPARQL query picked them up as two separate games.
Yugoslavia itself didn't even exist anymore in '97 but apparently the football team still used the name!
On this day (Oct 29):
Hungary 🇭🇺 1 : 7 Federal Republic of Yugoslavia (1997)
Hungary 🇭🇺 1 : 7 Serbia and Montenegro (1997)
Greece 🇬🇷 4 : 1 🇭🇺 Hungary (1978)
Hungary 🇭🇺 6 : 0 🇧🇴 Bolivia (1977)
German Democratic Republic 1 : 0 🇭🇺 Hungary (1967)
Hungary 🇭🇺 1 : 7 Federal Republic of Yugoslavia (1997)
Hungary 🇭🇺 1 : 7 Serbia and Montenegro (1997)
Greece 🇬🇷 4 : 1 🇭🇺 Hungary (1978)
Hungary 🇭🇺 6 : 0 🇧🇴 Bolivia (1977)
German Democratic Republic 1 : 0 🇭🇺 Hungary (1967)
October 30, 2025 at 1:50 PM
Another fun data problem with my bot.
The Wikipedia data has both Serbia & Montenegro AND Yugoslavia as opponents of the same game, so my SPARQL query picked them up as two separate games.
Yugoslavia itself didn't even exist anymore in '97 but apparently the football team still used the name!
The Wikipedia data has both Serbia & Montenegro AND Yugoslavia as opponents of the same game, so my SPARQL query picked them up as two separate games.
Yugoslavia itself didn't even exist anymore in '97 but apparently the football team still used the name!
Do you like any of the following: data science, real life data case studies, musings on our AI future, philosophy?
If you answered yes to any of the above, my talk at @databsconf.com might be of interest to you!
#databs #datasky
If you answered yes to any of the above, my talk at @databsconf.com might be of interest to you!
#databs #datasky
Highlighting #DataBS Conf talks:
Who _are_ we, as data practitioners? What is it we actually _do_ ?
@davidasboth.com hosts our late-night philosophy session, to explore the changing role of the data scientist in a world of AI:
www.youtube.com/watch?v=uCR1...
Who _are_ we, as data practitioners? What is it we actually _do_ ?
@davidasboth.com hosts our late-night philosophy session, to explore the changing role of the data scientist in a world of AI:
www.youtube.com/watch?v=uCR1...
DataBS 2025 - 11 - David Asboth - The Data Science Ship of Theseus
YouTube video by Data Behind the Scenes
www.youtube.com
October 21, 2025 at 8:51 AM
Do you like any of the following: data science, real life data case studies, musings on our AI future, philosophy?
If you answered yes to any of the above, my talk at @databsconf.com might be of interest to you!
#databs #datasky
If you answered yes to any of the above, my talk at @databsconf.com might be of interest to you!
#databs #datasky
This is beyond awful. Danya was a huge influence on me not even as a chess player but as an educator. His speedruns are a gold standard of teaching for me. Seemed like a genuinely lovely guy.
Someone somewhere commented "I just assumed I'd be learning from Danya my whole life, you know?" 💔
Someone somewhere commented "I just assumed I'd be learning from Danya my whole life, you know?" 💔
GM Daniel Naroditsky passed away. He was a talented chess player, commentator, and educator. FIDE extends its deepest condolences to Daniel’s family and loved ones.
October 20, 2025 at 8:16 PM
This is beyond awful. Danya was a huge influence on me not even as a chess player but as an educator. His speedruns are a gold standard of teaching for me. Seemed like a genuinely lovely guy.
Someone somewhere commented "I just assumed I'd be learning from Danya my whole life, you know?" 💔
Someone somewhere commented "I just assumed I'd be learning from Danya my whole life, you know?" 💔
#dataBS I need to hear your favourite dataset you've used for teaching data topics.
My criteria:
- smallish, in the order of 10k rows, not millions
- wide enough with a mix of data types to allow open-ended questions
- not seen in your typical online data tutorials
🚫No ocean liners🚫
My criteria:
- smallish, in the order of 10k rows, not millions
- wide enough with a mix of data types to allow open-ended questions
- not seen in your typical online data tutorials
🚫No ocean liners🚫
October 17, 2025 at 1:20 PM
#dataBS I need to hear your favourite dataset you've used for teaching data topics.
My criteria:
- smallish, in the order of 10k rows, not millions
- wide enough with a mix of data types to allow open-ended questions
- not seen in your typical online data tutorials
🚫No ocean liners🚫
My criteria:
- smallish, in the order of 10k rows, not millions
- wide enough with a mix of data types to allow open-ended questions
- not seen in your typical online data tutorials
🚫No ocean liners🚫
Today's #databs mishap. Tableau can automatically convert strings to dates, right? Turns out if your date is in the form "Jan-88", then the ONLY short code for September that Tableau will accept is "Sept", NOT "Sep".
This explains all the NULL values and gaps in my time series 😭
#datasky
This explains all the NULL values and gaps in my time series 😭
#datasky
October 16, 2025 at 1:17 PM
Reposted by David Asboth
Apparently there are a bunch of new people coming over from Twitter... Drop your data science starter packs in the replies for people to follow!
October 16, 2025 at 3:23 AM
Apparently there are a bunch of new people coming over from Twitter... Drop your data science starter packs in the replies for people to follow!
"Time to browse LinkedIn"
(not limited to my profession of course)
(not limited to my profession of course)
In honor of spooky month, share a 4 word horror story that only someone in your profession would understand
I'll go first: Six page commercial lease.
I'll go first: Six page commercial lease.
October 12, 2025 at 6:54 PM
"Time to browse LinkedIn"
(not limited to my profession of course)
(not limited to my profession of course)
The problem with having ideas, writing them down, and not acting on them for ages is that I have absolutely no idea what I was thinking.
I have a blog post idea written down that just says: learning "experiences"
What does that mean?????
I have a blog post idea written down that just says: learning "experiences"
What does that mean?????
October 11, 2025 at 6:39 PM
The problem with having ideas, writing them down, and not acting on them for ages is that I have absolutely no idea what I was thinking.
I have a blog post idea written down that just says: learning "experiences"
What does that mean?????
I have a blog post idea written down that just says: learning "experiences"
What does that mean?????
Worth reading this deeper dive into how well GenAI tools do with dataviz problems, by @nrennie.bsky.social
#databs #datasky
#databs #datasky
Wondering if you can outsource your data viz work to ChatGPT? 📊
I tested out a few different generative AI tools, giving them prompts to visualise two different data sets. If you're interested in the results, you can read them here: nrennie.rbind.io/blog/gen-ai-...
#RStats #Python #DataViz #GenAI
I tested out a few different generative AI tools, giving them prompts to visualise two different data sets. If you're interested in the results, you can read them here: nrennie.rbind.io/blog/gen-ai-...
#RStats #Python #DataViz #GenAI
Generative AI for Data Visualisation – Nicola Rennie
Can generative AI create good data visualisations? This blog post compares the performance of ChatGPT, Claude, Copilot, and Gemini when presented with a generic request to visualise a dataset.
nrennie.rbind.io
October 9, 2025 at 10:16 AM
Worth reading this deeper dive into how well GenAI tools do with dataviz problems, by @nrennie.bsky.social
#databs #datasky
#databs #datasky
Reposted by David Asboth
The videos from the #DataBS Conference have made it to YouTube!
Over the coming days, we'll highlight sessisons from the event -- one post per talk.
You can also sneak ahead by going straight to the YouTube playlist, here:
www.youtube.com/playlist?lis...
Over the coming days, we'll highlight sessisons from the event -- one post per talk.
You can also sneak ahead by going straight to the YouTube playlist, here:
www.youtube.com/playlist?lis...
Data Behind the Scenes 2025 - YouTube
Data Behind the Scenes is a data conference by data practitioners, for data practitioners. We come together to share stories and experiences about the the me...
www.youtube.com
October 6, 2025 at 5:15 PM
The videos from the #DataBS Conference have made it to YouTube!
Over the coming days, we'll highlight sessisons from the event -- one post per talk.
You can also sneak ahead by going straight to the YouTube playlist, here:
www.youtube.com/playlist?lis...
Over the coming days, we'll highlight sessisons from the event -- one post per talk.
You can also sneak ahead by going straight to the YouTube playlist, here:
www.youtube.com/playlist?lis...
I got to round 2 of this challenge so I'm writing my second short story this weekend!
I only joined to force myself to write a bit more and I'm so happy I get to write another!
I'll post all the stories I write for this when the organisers say it's OK (a few weeks before I can post the first one).
I only joined to force myself to write a bit more and I'm so happy I get to write another!
I'll post all the stories I write for this when the organisers say it's OK (a few weeks before I can post the first one).
I like writing and want to do more so I'll be taking a little writing challenge later this year.
Join me in the @nycmidnight.bsky.social 500-word Fiction Challenge on August 15th!
Learn more at nycmidnight.com/500
Join me in the @nycmidnight.bsky.social 500-word Fiction Challenge on August 15th!
Learn more at nycmidnight.com/500
October 4, 2025 at 12:04 PM
I got to round 2 of this challenge so I'm writing my second short story this weekend!
I only joined to force myself to write a bit more and I'm so happy I get to write another!
I'll post all the stories I write for this when the organisers say it's OK (a few weeks before I can post the first one).
I only joined to force myself to write a bit more and I'm so happy I get to write another!
I'll post all the stories I write for this when the organisers say it's OK (a few weeks before I can post the first one).
So many questions. "In the same way vibe coding has transformed software development".
Has it?
Do we want spreadsheets, on which entire companies are run, being amended by something allegedly less than 60% accurate?!?
Have these companies just stopped pretending like they're launching useful AI?
Has it?
Do we want spreadsheets, on which entire companies are run, being amended by something allegedly less than 60% accurate?!?
Have these companies just stopped pretending like they're launching useful AI?
October 1, 2025 at 8:16 PM
So many questions. "In the same way vibe coding has transformed software development".
Has it?
Do we want spreadsheets, on which entire companies are run, being amended by something allegedly less than 60% accurate?!?
Have these companies just stopped pretending like they're launching useful AI?
Has it?
Do we want spreadsheets, on which entire companies are run, being amended by something allegedly less than 60% accurate?!?
Have these companies just stopped pretending like they're launching useful AI?
This is frankly the problem with practically every digital project ever conceived. Today the discussion is about digital IDs in the UK which sound like a NIGHTMARE from a technical perspective before we even consider all the other implications. No way it ends well.
the implementation will not be well thought through or correct though.
September 26, 2025 at 4:17 PM
This is frankly the problem with practically every digital project ever conceived. Today the discussion is about digital IDs in the UK which sound like a NIGHTMARE from a technical perspective before we even consider all the other implications. No way it ends well.