Kevin Yang
@yang3kc.bsky.social
Kaicheng Yang, PhD | Assistant professor @Binghamton University CS | Computational social science, bots, + | views mine
Our introduction to the National Internet Observatory, a platform for collecting and sharing online activity data designed for researchers, is happening this afternoon at #ica25
June 15, 2025 at 4:56 PM
Our introduction to the National Internet Observatory, a platform for collecting and sharing online activity data designed for researchers, is happening this afternoon at #ica25
Specifically, we identify a number of obstacles researchers face 🔒:
- Lack of awareness/clarity
- Complex/burdensome application processes
- Arbitrary/opaque rejections
- Unreliable/low-quality infra and data
- Exclusion of non-academic and Global South researchers
2/3
- Lack of awareness/clarity
- Complex/burdensome application processes
- Arbitrary/opaque rejections
- Unreliable/low-quality infra and data
- Exclusion of non-academic and Global South researchers
2/3
May 19, 2025 at 2:11 PM
Specifically, we identify a number of obstacles researchers face 🔒:
- Lack of awareness/clarity
- Complex/burdensome application processes
- Arbitrary/opaque rejections
- Unreliable/low-quality infra and data
- Exclusion of non-academic and Global South researchers
2/3
- Lack of awareness/clarity
- Complex/burdensome application processes
- Arbitrary/opaque rejections
- Unreliable/low-quality infra and data
- Exclusion of non-academic and Global South researchers
2/3
[New WP] With the closure of major social media APIs and the new data access mandates under DSA, we enter what we call the "post-post-API" era. But have researchers obtained the data they need? Our recent survey (180) + interview (19) study suggests a stark reality.
🔗 arxiv.org/abs/2505.09877
1/3
🔗 arxiv.org/abs/2505.09877
1/3
May 19, 2025 at 2:11 PM
[New WP] With the closure of major social media APIs and the new data access mandates under DSA, we enter what we call the "post-post-API" era. But have researchers obtained the data they need? Our recent survey (180) + interview (19) study suggests a stark reality.
🔗 arxiv.org/abs/2505.09877
1/3
🔗 arxiv.org/abs/2505.09877
1/3
At this rate, I'm going to review 100+ papers this year.
February 6, 2025 at 1:55 AM
At this rate, I'm going to review 100+ papers this year.
We derive the data from a panel of over 1.5M Twitter users matched against their US registration records.
📰 Preprint: arxiv.org/abs/2501.09035
🧑💻 Github (with data): github.com/LazerLab/Dom...
🖥️ Interactive app: domaindemoexplorer.streamlit.app
📻 Podcast: notebooklm.google.com/notebook/e27...
2/3
📰 Preprint: arxiv.org/abs/2501.09035
🧑💻 Github (with data): github.com/LazerLab/Dom...
🖥️ Interactive app: domaindemoexplorer.streamlit.app
📻 Podcast: notebooklm.google.com/notebook/e27...
2/3
January 17, 2025 at 3:40 PM
We derive the data from a panel of over 1.5M Twitter users matched against their US registration records.
📰 Preprint: arxiv.org/abs/2501.09035
🧑💻 Github (with data): github.com/LazerLab/Dom...
🖥️ Interactive app: domaindemoexplorer.streamlit.app
📻 Podcast: notebooklm.google.com/notebook/e27...
2/3
📰 Preprint: arxiv.org/abs/2501.09035
🧑💻 Github (with data): github.com/LazerLab/Dom...
🖥️ Interactive app: domaindemoexplorer.streamlit.app
📻 Podcast: notebooklm.google.com/notebook/e27...
2/3
Introducing “DomainDemo: a dataset of domain-sharing activities among different demographic groups on Twitter.”
Today, we release five derived metrics of over 129,000 domains, quantifying their characteristics such as geographical reach and audience partisanship.
1/3
Today, we release five derived metrics of over 129,000 domains, quantifying their characteristics such as geographical reach and audience partisanship.
1/3
January 17, 2025 at 3:40 PM
Introducing “DomainDemo: a dataset of domain-sharing activities among different demographic groups on Twitter.”
Today, we release five derived metrics of over 129,000 domains, quantifying their characteristics such as geographical reach and audience partisanship.
1/3
Today, we release five derived metrics of over 129,000 domains, quantifying their characteristics such as geographical reach and audience partisanship.
1/3
Might be too much though 😅
December 7, 2024 at 2:37 AM
Might be too much though 😅
Great to see this work about information operation finally out! I think the coolest part is the control data added by the team, which makes all sorts of analyses and applications possible.
Link: arxiv.org/abs/2411.10609
Link: arxiv.org/abs/2411.10609
November 19, 2024 at 3:23 PM
Great to see this work about information operation finally out! I think the coolest part is the control data added by the team, which makes all sorts of analyses and applications possible.
Link: arxiv.org/abs/2411.10609
Link: arxiv.org/abs/2411.10609
4. LLM to generate git commits
Another small tool I built. It can read the diff info and produce a few candidate git commit information for you to choose from.
Link: github.com/yang3kc/llm_...
Another small tool I built. It can read the diff info and produce a few candidate git commit information for you to choose from.
Link: github.com/yang3kc/llm_...
November 18, 2024 at 3:55 PM
4. LLM to generate git commits
Another small tool I built. It can read the diff info and produce a few candidate git commit information for you to choose from.
Link: github.com/yang3kc/llm_...
Another small tool I built. It can read the diff info and produce a few candidate git commit information for you to choose from.
Link: github.com/yang3kc/llm_...
3. LLM to filter relevant arxiv papers
This is a tool I built myself. It can download the new papers from arxiv then use an LLM determine their relevance to the topics of your interests. Add a web interface lately, so it's easier to use.
Link: github.com/yang3kc/dail...
This is a tool I built myself. It can download the new papers from arxiv then use an LLM determine their relevance to the topics of your interests. Add a web interface lately, so it's easier to use.
Link: github.com/yang3kc/dail...
November 18, 2024 at 3:55 PM
3. LLM to filter relevant arxiv papers
This is a tool I built myself. It can download the new papers from arxiv then use an LLM determine their relevance to the topics of your interests. Add a web interface lately, so it's easier to use.
Link: github.com/yang3kc/dail...
This is a tool I built myself. It can download the new papers from arxiv then use an LLM determine their relevance to the topics of your interests. Add a web interface lately, so it's easier to use.
Link: github.com/yang3kc/dail...
Super interesting findings on when AI+human is good or bad. Consistent with my own experience. Still critical to learn and practice in the age of AI.
Link: www.nature.com/articles/s41...
Link: www.nature.com/articles/s41...
November 18, 2024 at 3:30 PM
Super interesting findings on when AI+human is good or bad. Consistent with my own experience. Still critical to learn and practice in the age of AI.
Link: www.nature.com/articles/s41...
Link: www.nature.com/articles/s41...
TIL that you can ignore the .gitignore file in the .gitignore file 🤯
November 17, 2024 at 2:21 AM
TIL that you can ignore the .gitignore file in the .gitignore file 🤯
Have been using Cursor for writing LaTeX documents for a while. The idea was simple: Cursor is helpful for coding, and TeX is code in a way, so I gave it a try, and it worked perfectly.
It does require some configuration, so I'm sharing mine here: github.com/yang3kc/curs....
1/2
It does require some configuration, so I'm sharing mine here: github.com/yang3kc/curs....
1/2
November 13, 2024 at 1:10 PM
Have been using Cursor for writing LaTeX documents for a while. The idea was simple: Cursor is helpful for coding, and TeX is code in a way, so I gave it a try, and it worked perfectly.
It does require some configuration, so I'm sharing mine here: github.com/yang3kc/curs....
1/2
It does require some configuration, so I'm sharing mine here: github.com/yang3kc/curs....
1/2
New blog on prompt engineering for data analysis. I think many existing studies are not rigorous enough in terms of prompt design. But there are ways to improve.
Link: open.substack.com/pub/yang3kc/...
Link: open.substack.com/pub/yang3kc/...
January 24, 2024 at 5:01 PM
New blog on prompt engineering for data analysis. I think many existing studies are not rigorous enough in terms of prompt design. But there are ways to improve.
Link: open.substack.com/pub/yang3kc/...
Link: open.substack.com/pub/yang3kc/...
December 5, 2023 at 1:12 PM
Photo touching up in the age of AI 😎
October 19, 2023 at 2:28 PM
Photo touching up in the age of AI 😎
People used to joke about #ICWSM being a Twitter conference, so I analyzed the papers to see if it's true. And yes, over 30% of the papers mentioned "twitter." I might have contributed a few times . Wonder where the community is headed without free data .
October 12, 2023 at 5:43 PM
People used to joke about #ICWSM being a Twitter conference, so I analyzed the papers to see if it's true. And yes, over 30% of the papers mentioned "twitter." I might have contributed a few times . Wonder where the community is headed without free data .
The Bluesky icon looks wired when added as an app on macOS Sonoma
October 2, 2023 at 2:42 AM
The Bluesky icon looks wired when added as an app on macOS Sonoma