@uwdata

Visualization & data analysis at @uw @uwcse. In a previous life was the Stanford Vis Group.

 Seattle, WA
idl.cs.washington.edu
Joined September 2013
  • Tweets516
  • Following276
  • Followers7.4K
30Media
Thanks to all who joined us for our @ieeevis Seattle satellite! Come meet some of the participants and learn about their work: youtube.com/watch?v=EgBQo5QX…
1
3
3
20
With DracoGPT, Will Wang shows how to extract and model visualization design preferences from generative AI systems — enabling new ways to quantify, evaluate, and efficiently reuse LLM-based chart recommendations. #ieeevis idl.uw.edu/papers/dracogpt

DracoGPT: Extracting Visualization Design Preferences from Large Language Models

idl.uw.edu
1
3
18
2,718
Congratulations to IDL alum @domoritz for winning a VGTC Significant New Researcher award!! A premier honor for early career researchers in visualization!
7
7
4
59
9,540
Interactive Data Lab retweeted
Eduardo Blancas@edublancas2 Oct 2024
Mosaic is the future of data visualization. 10M cross-filtered rows with client-side processing. Powered by @duckdb
5
10
1
139
13,882
Chaining LLM calls can improve output quality, but navigating the massive space of task decompositions is challenging. Revisiting the established field of crowdsourcing, we distill strategies for effective LLM chain design and identify opportunities for future research. [1/11]
1
11
2
51
7,563
Interactive Data Lab retweeted
Michelle Lam@michelle123lam18 Apr 2024
“Can we get a new text analysis tool?” “No—we have Topic Model at home” Topic Model at home: outputs vague keywords; needs constant parameter fiddling🫠 Is there a better way? We introduce LLooM, a concept induction tool to explore text data in terms of interpretable concepts🧵
4
39
7
203
38,294
Interactive Data Lab retweeted
Vega, Vega-Lite & Altair@vega_vis4 Apr 2024
The Vega Project is happy to announce the release of version 5.3.0 of the Vega-Altair Python visualization library. This release has been 4 months in the making and includes enhancements, fixes, and documentation improvements from 11 contributors. Highlights in 🧵
2
16
2
68
9,496
Interactive Data Lab retweeted
Mike Bostock@mbostock14 Mar 2024
Observable Framework 1.3 🆕 integrates @uwdata’s Mosaic vgplot, which can concisely expressive performant coordinated views of millions of data points. observablehq.com/framework/l…
5
35
5
270
21,207
Interactive Data Lab retweeted
Jeffrey Heer@jeffrey_heer27 Feb 2024
Interact with millions of data points in real-time with Mosaic, now with support for geospatial data. Exploring 1M taxi pickups and dropoffs in NYC:
6
39
5
339
30,431
Chaining LLMs together to overcome LLM errors is an emerging, yet challenging, technique. What can we learn from crowdsourcing, which has long dealt with the challenge of decomposing complex work? We delve into this question in our new preprint: arxiv.org/abs/2312.11681 [1/9]
3
14
4
84
13,151
Interactive Data Lab retweeted
Allen School@uwcse11 Dec 2023
As a kid, #UWAllen @uwdata professor & alum Leilani Battle wanted to be a game developer. At @UW, instead of games, she immersed herself in data. She earned the 2023 @ieeevgtc Visualization Significant New Researcher Award for her work. #ThisIsUW #dataviz news.cs.washington.edu/2023/…

Not playing around: Allen School’s Leilani Battle receives VGTC Visualization Significant New...

Allen School professor and alum Leilani Battle (B.S., ‘11) originally wanted to be a game developer. As a kid growing up in Bremerton Washington, Battle saw a glimpse of her future every time she...

news.cs.washington.edu
3
16
2,118
Interactive Data Lab retweeted
IEEE VIS@ieeevis25 Oct 2023
Join us in congratulating Dr. Leilani Battle for receiving the 2023 VGTC Visualization Significant New Researcher Award at #IEEEVIS. 🏆 👏 @leibatt is honored for her work showing how data exploration systems can slow down, confuse, bias, and even mislead analysts.
1
8
1
74
5,346
Congrats to IDL co-director Leilani Battle (@leibatt) for the 2023 VGTC Significant New Researcher Award! #IEEEVIS2023 ieeecs-media.computer.org/tc…
2
2
29
2,549
And congrats also to the other Significant New Researcher recipient, UW alum Matt Kay (@mjskay)!
1
9
1,499
Interactive Data Lab retweeted
Jared Wilber@jdwlbr19 Sep 2023
Really enjoying playing with @uwdata 's Mosaic, linking @duckdb with @observablehq plot feels so natural and responsive - just a joy
3
27
1
189
15,851
Interactive Data Lab retweeted
Arvind Satyanarayan@arvindsatya130 Jun 2023
VisText has been a *ton* of work, and 2 yrs of solid effort. So it's exciting to (finally!) be able to talk about it, and it's gratifying to see it featured on @MIT's homepage. Lead author, @bennyjtang, has a great thread w/details below 👇 And I wanted to offer a few thoughts
A screenshot of the homepage of mit.edu that features a two column layout: a sidebar on the left hand side (with a search box and links to top resources), and then the main feature on the right hand side. This feature includes an image of a bar chart (where each bar is labeled with one letter from the word "caption"). Underneath the image is text that reads "Researchers have developed a dataset for teaching an AI to write better chart captions, improving accessibility for blind and low-vision individuals. A model trained with this dataset could translate “‘chart language’ to English,” Angie Boggust explains."

ALT A screenshot of the homepage of mit.edu that features a two column layout: a sidebar on the left hand side (with a search box and links to top resources), and then the main feature on the right hand side. This feature includes an image of a bar chart (where each bar is labeled with one letter from the word "caption"). Underneath the image is text that reads "Researchers have developed a dataset for teaching an AI to write better chart captions, improving accessibility for blind and low-vision individuals. A model trained with this dataset could translate “‘chart language’ to English,” Angie Boggust explains."

Ben Tang@bennyjtang30 Jun 2023
Chart captioning is hard, both for humans & AI. Today, we’re introducing VisText: a benchmark dataset of 12k+ visually-diverse charts w/ rich captions for automatic captioning (w/ @angie_boggust @arvindsatya1) 📄: vis.csail.mit.edu/pubs/viste… 💻: github.com/mitvis/vistext #ACL2023NLP
A montage of various different bar, line, and area charts across a mix of visual styles to illustrate the diversity of the VisText dataset.

ALT A montage of various different bar, line, and area charts across a mix of visual styles to illustrate the diversity of the VisText dataset.

2
27
1
119
25,552
Interactive Data Lab retweeted
Hamilton Ulmer
@hamiltonulmer21 Jun 2023
Mosaic looks really great! Great to see a viz library thread down to the db level. DuckDB is such a natural choice for something like this. uwdata.github.io/mosaic/
2
66
2
450
58,154
New project! Mosaic is a framework for scalable and interoperable visualization that unifies visualization toolkits with scalable visualization methods and query optimizations. uwdata.github.io/mosaic/what…
2
42
13
174
52,247
more replies
To test Mosaic we developed vgplot, a Mosaic-based grammar of interactive graphics that uses Observable Plot to perform chart rendering. Our initial benchmark studies show order-of-magnitude performance improvements over Vega, VegaFusion, and standard Observable Plot.
Charts of visualization rendering performance, comparing different tools (Vega, VegaFusion, Observable Plot, and Mosaic) and chart types (average bars, linear regression, 2d histograms, area charts, density contours, and hexbins). Mosaic provides superior performance, often by one or more orders of magnitude.

ALT Charts of visualization rendering performance, comparing different tools (Vega, VegaFusion, Observable Plot, and Mosaic) and chart types (average bars, linear regression, 2d histograms, area charts, density contours, and hexbins). Mosaic provides superior performance, often by one or more orders of magnitude.

1
1
14
2,525
While Mosaic is still "research grade" software, we think the approach is quite promising. Feel free to share feedback or contribute on GitHub! github.com/uwdata/mosaic

GitHub - uwdata/mosaic: An extensible framework for linking databases and interactive views.

An extensible framework for linking databases and interactive views. - uwdata/mosaic

github.com
1
1
14
1,947
Load more