The Global Carbon Budget 2026 was published today.
Published today in Earth System Science Data: The Global Carbon Budget 2026
essd.copernicus.org/preprints/es...
Published today in Earth System Science Data: The Global Carbon Budget 2026
essd.copernicus.org/preprints/es...
November 10, 2026 at 4:02 AM
The Global Carbon Budget 2026 was published today.
Published today in Earth System Science Data: The Global Carbon Budget 2026
essd.copernicus.org/preprints/es...
Published today in Earth System Science Data: The Global Carbon Budget 2026
essd.copernicus.org/preprints/es...
when there are two of a thing it means that exactly 50% of people choose each side. that's just data science.
November 8, 2026 at 3:12 AM
when there are two of a thing it means that exactly 50% of people choose each side. that's just data science.
Yes, as if designing an experiment and collecting data with intent doesn’t matter.
The same rock 90% of “data science” and “big data” use cases foundered on
The same rock 90% of “data science” and “big data” use cases foundered on
November 8, 2026 at 10:03 PM
Yes, as if designing an experiment and collecting data with intent doesn’t matter.
The same rock 90% of “data science” and “big data” use cases foundered on
The same rock 90% of “data science” and “big data” use cases foundered on
New manuscript
Understanding Pedagogical Content Knowledge of Introductory Data Science Instructors: An Inaugural Framework
Published in Teaching and Teacher Education
www.sciencedirect.com/science/arti...
#datascienceed
Understanding Pedagogical Content Knowledge of Introductory Data Science Instructors: An Inaugural Framework
Published in Teaching and Teacher Education
www.sciencedirect.com/science/arti...
#datascienceed
Understanding Pedagogical Content Knowledge of Introductory Data Science Instructors: An Inaugural Framework
As data science emerges as a distinct academic discipline, introductory data science (IDS) courses play a key role in shaping students’ foundational u…
www.sciencedirect.com
November 9, 2026 at 9:24 PM
New manuscript
Understanding Pedagogical Content Knowledge of Introductory Data Science Instructors: An Inaugural Framework
Published in Teaching and Teacher Education
www.sciencedirect.com/science/arti...
#datascienceed
Understanding Pedagogical Content Knowledge of Introductory Data Science Instructors: An Inaugural Framework
Published in Teaching and Teacher Education
www.sciencedirect.com/science/arti...
#datascienceed
This isn't how science works. You're attempting to prove your assumption with inapplicable data.
November 10, 2026 at 3:03 PM
This isn't how science works. You're attempting to prove your assumption with inapplicable data.
I'm a grumpy old "sports hacks" cynic. Really, almost nothing works (or works well). Stick to science-informed basics❗
3 Sports Science Hacks That Don’t Work 👇
New data on ketones, laser therapy, and hot/cold recovery baths dial back the hype.
www.outsideonline.com/health/train...
3 Sports Science Hacks That Don’t Work 👇
New data on ketones, laser therapy, and hot/cold recovery baths dial back the hype.
www.outsideonline.com/health/train...
3 Popular Recovery Trends That Just Failed in the Lab
New data on ketones, laser therapy, and hot/cold recovery baths dial back the hype.
www.outsideonline.com
November 9, 2026 at 6:13 PM
I'm a grumpy old "sports hacks" cynic. Really, almost nothing works (or works well). Stick to science-informed basics❗
3 Sports Science Hacks That Don’t Work 👇
New data on ketones, laser therapy, and hot/cold recovery baths dial back the hype.
www.outsideonline.com/health/train...
3 Sports Science Hacks That Don’t Work 👇
New data on ketones, laser therapy, and hot/cold recovery baths dial back the hype.
www.outsideonline.com/health/train...
Critical point by @cailinmeister.bsky.social.. academics coming out saying they weren’t paid by meta perpetuates a myth about how industry influences science. The reality is much more pernicious; everything from selective data access to unrestricted grants.
Meta is trying to discredit researchers who argue their products are harmful. This does not mean that researchers “on their side” are corrupt, and focusing on corruption in science in this case is to entirely miss what is going wrong.
Industry almost never buys corrupt researchers, and this is really important
Conflating industry influence in science with researcher bias drives repeated failures to protect science
cailinoconnor.substack.com
November 9, 2026 at 9:12 PM
Critical point by @cailinmeister.bsky.social.. academics coming out saying they weren’t paid by meta perpetuates a myth about how industry influences science. The reality is much more pernicious; everything from selective data access to unrestricted grants.
Join us in advancing data science and AI research! The Johns Hopkins Data Science and AI Institute Postdoctoral Fellowship Program is now accepting applications for the 2027–2028 academic year. Apply now! Deadline: Dec. 11, 2026.
Details and apply: apply.interfolio.com/196237
Details and apply: apply.interfolio.com/196237
November 9, 2026 at 4:01 PM
Join us in advancing data science and AI research! The Johns Hopkins Data Science and AI Institute Postdoctoral Fellowship Program is now accepting applications for the 2027–2028 academic year. Apply now! Deadline: Dec. 11, 2026.
Details and apply: apply.interfolio.com/196237
Details and apply: apply.interfolio.com/196237
9 Ways Public Doubt About Long COVID Is Manufactured — Using Tactics Similar to Big Tobacco and Climate Science Denial.
Source: x.com/SalvMattera/...
Source: x.com/SalvMattera/...
November 9, 2026 at 6:24 AM
9 Ways Public Doubt About Long COVID Is Manufactured — Using Tactics Similar to Big Tobacco and Climate Science Denial.
Source: x.com/SalvMattera/...
Source: x.com/SalvMattera/...
I mean, it *is* data science 101, but have you considered "yeah but I want it to"
November 10, 2026 at 4:19 PM
I mean, it *is* data science 101, but have you considered "yeah but I want it to"
The first lesson of data is it’s just data. You can’t extrapolate causation without actual tests and an election is just not a good test! It’s useful for an election to make models sure, but that doesn’t mean you ever know why! It should be data science 101!!
November 10, 2026 at 4:12 PM
The first lesson of data is it’s just data. You can’t extrapolate causation without actual tests and an election is just not a good test! It’s useful for an election to make models sure, but that doesn’t mean you ever know why! It should be data science 101!!
Wearable device makers are harvesting brain data to train AI—without telling users | Science | AAAS www.science.org/content/arti...
Wearable device makers are harvesting brain data to train AI—without telling users
Science investigation reveals firms have amassed tens of millions of hours of brain wave data from neurotech headbands, and face few constraints in using it
www.science.org
November 7, 2026 at 2:06 PM
Wearable device makers are harvesting brain data to train AI—without telling users | Science | AAAS www.science.org/content/arti...
JHU Data Science and AI Institute Postdoctoral Fellowship program is searching for postdocs (due Dec 11th).
If you're interested in working with me on computational social science, NLP, and AI for human behavior & society, apply here: apply.interfolio.com/196237
If you're interested in working with me on computational social science, NLP, and AI for human behavior & society, apply here: apply.interfolio.com/196237
Apply - Interfolio
{{$ctrl.$state.data.pageTitle}} - Apply - Interfolio
apply.interfolio.com
November 9, 2026 at 8:04 PM
JHU Data Science and AI Institute Postdoctoral Fellowship program is searching for postdocs (due Dec 11th).
If you're interested in working with me on computational social science, NLP, and AI for human behavior & society, apply here: apply.interfolio.com/196237
If you're interested in working with me on computational social science, NLP, and AI for human behavior & society, apply here: apply.interfolio.com/196237
Some cool data science that isn't deep learning or LLMs for a change: e-means, a new algorithm for k-means clustering that fits faster and is less prone to getting trapped in local minima. Has some nice animated data vis too.
e-means.github.io
e-means.github.io
e-means: 10x faster k-means at equal loss - Blog
Modern optimization and scaling laws for k-means clustering
e-means.github.io
November 10, 2026 at 12:52 PM
Some cool data science that isn't deep learning or LLMs for a change: e-means, a new algorithm for k-means clustering that fits faster and is less prone to getting trapped in local minima. Has some nice animated data vis too.
e-means.github.io
e-means.github.io
Science, stop making AI Data Centers! We need lithium ion batteries that last more than a year.
November 10, 2026 at 12:28 AM
Science, stop making AI Data Centers! We need lithium ion batteries that last more than a year.
The core appeal of “popularism” was always the idea that liberals could win majorities while doing small-ball policy work. The polling and data science aspects were just window dressing and it had obvious but rarely acknowledged blind spots on earned media and salience.
Many of the political predictions in this hagiographic 2021 Ezra Klein profile of David Shor have aged like cottage cheese. www.nytimes.com/2021/10/08/o...
November 8, 2026 at 1:22 PM
The core appeal of “popularism” was always the idea that liberals could win majorities while doing small-ball policy work. The polling and data science aspects were just window dressing and it had obvious but rarely acknowledged blind spots on earned media and salience.
I was pretty sure that science didn't work by taking statistically similar science words and jamming them together in a semantically correct way but it's nice to have data
"Executives leading the companies that discover new drugs say AI isn’t going to cure cancer anytime soon. They are dimming the hype around the technology’s use in drug research, saying human biology is too complicated even for superintelligence." www.wsj.com/tech/ai/sili...
Silicon Valley Promises AI Will Cure Disease. Drugmaker CEOs Aren’t Buying It.
No drug designed by artificial intelligence has won FDA clearance, and pharma leaders say biology is the obstacle.
www.wsj.com
November 8, 2026 at 4:43 PM
I was pretty sure that science didn't work by taking statistically similar science words and jamming them together in a semantically correct way but it's nice to have data
Have you ever seen satellite images and wondered how machine learning models know what’s on the ground?
Tomorrow, @dsollberger.bsky.social will show us an end-to-end GeoAI classification workflow on the Data Science Lab!
Derek, thank you for the classic Sim City vibes 🏙️
Join live: pos.it/dslab
Tomorrow, @dsollberger.bsky.social will show us an end-to-end GeoAI classification workflow on the Data Science Lab!
Derek, thank you for the classic Sim City vibes 🏙️
Join live: pos.it/dslab
November 9, 2026 at 3:49 PM
Have you ever seen satellite images and wondered how machine learning models know what’s on the ground?
Tomorrow, @dsollberger.bsky.social will show us an end-to-end GeoAI classification workflow on the Data Science Lab!
Derek, thank you for the classic Sim City vibes 🏙️
Join live: pos.it/dslab
Tomorrow, @dsollberger.bsky.social will show us an end-to-end GeoAI classification workflow on the Data Science Lab!
Derek, thank you for the classic Sim City vibes 🏙️
Join live: pos.it/dslab
The question I want answered: How many data centres do we need to conduct science and analyse health data, etc? How much of what we are getting is to produce entertainment and tat? Because we can so get on fine without the latter.
Zap Hosting Data Center Server Racks
Alt: Zap Hosting Data Center Server Racks
static.klipy.com
November 7, 2026 at 9:48 AM
The question I want answered: How many data centres do we need to conduct science and analyse health data, etc? How much of what we are getting is to produce entertainment and tat? Because we can so get on fine without the latter.
Companies are using wearable neurotechnologies to train #AI models of the brain—without telling users.
Learn more: https://scim.ag/3Ux7r7h @pulitzercenter.org
Learn more: https://scim.ag/3Ux7r7h @pulitzercenter.org
November 6, 2026 at 11:30 PM
Companies are using wearable neurotechnologies to train #AI models of the brain—without telling users.
Learn more: https://scim.ag/3Ux7r7h @pulitzercenter.org
Learn more: https://scim.ag/3Ux7r7h @pulitzercenter.org
🤖 AI agents curate science data
On 14 biomedical and chemistry tasks, AI built on agent-curated data scored 86.30% on medical questions against 80.64% for expert-curated data. Over 12 weeks of manual work took days.
🔗 doi.org/10.1038/s41467-026-78887-x
#AI #Science #Chemistry 🧪
On 14 biomedical and chemistry tasks, AI built on agent-curated data scored 86.30% on medical questions against 80.64% for expert-curated data. Over 12 weeks of manual work took days.
🔗 doi.org/10.1038/s41467-026-78887-x
#AI #Science #Chemistry 🧪
Agentic intelligent data engine for scientific large language models - Nature Communications
Here, authors develop AIDE, a multiagent framework that automates scientific data curation, and demonstrate across biomedical and chemical tasks that its automatically curated datasets can match or ou...
doi.org
November 7, 2026 at 12:36 PM
🤖 AI agents curate science data
On 14 biomedical and chemistry tasks, AI built on agent-curated data scored 86.30% on medical questions against 80.64% for expert-curated data. Over 12 weeks of manual work took days.
🔗 doi.org/10.1038/s41467-026-78887-x
#AI #Science #Chemistry 🧪
On 14 biomedical and chemistry tasks, AI built on agent-curated data scored 86.30% on medical questions against 80.64% for expert-curated data. Over 12 weeks of manual work took days.
🔗 doi.org/10.1038/s41467-026-78887-x
#AI #Science #Chemistry 🧪
data schemas and information structures with well defined attributes building into a system model that everyone understands is STILL necessary and LLM’s won’t fix the problem of abandoning information architecture as a multidisciplinary cross functional art of human factors and computer science.
November 8, 2026 at 1:39 PM
data schemas and information structures with well defined attributes building into a system model that everyone understands is STILL necessary and LLM’s won’t fix the problem of abandoning information architecture as a multidisciplinary cross functional art of human factors and computer science.
For researchers already working with 2026 election data and looking for comparisons, don't forget about our back catalog of quality-assured, standardized national datasets with precinct-level results from past elections, including the 2022 and 2018 midterms.
MIT Election Data + Science Lab
This is the official Dataverse of the MIT Election Data and Science Lab (MEDSL).
buff.ly
November 5, 2026 at 2:02 PM
For researchers already working with 2026 election data and looking for comparisons, don't forget about our back catalog of quality-assured, standardized national datasets with precinct-level results from past elections, including the 2022 and 2018 midterms.