Re: Web Charts

Civic Technology Community Group,

A quick update: I revised the Web Charts proposal. It is now a W3C WICG proposal: https://github.com/WICG/proposals/issues/97 .

For new members and participants: we have been discussing that, with AI systems: (1) end-users could engage in dialogues to search for, create, and style data visualizations [1], and (2) data visualizations could be generated and provided as components of AI systems' multimodal responses [2]. We have been discussing these topics in the contexts of civic technology, open government, and AI. We have been discussing the multimodal conversational search and exploration of large-scale public-sector data.

Please do feel free to share any ideas, comments, questions, or feedback about Web Charts and the above discussion topics in this mailing list or the WICG proposal!


Best regards,
Adam

[1] https://microsoft.github.io/lida/
[2] https://blogs.bing.com/search/may_2023/Bing-Preview-Release-Notes-Chat-History,-Charts,-Export,-and-more

P.S.: Also, with respect to Mike's points about misleading statistics and graphs, I found the following Wikipedia articles to be useful to learn more about his concerns:

  1.  https://en.wikipedia.org/wiki/Misuse_of_statistics
  2.  https://en.wikipedia.org/wiki/Misleading_graph

I am optimistic (in particular with Web Charts that provide separate input data, reusable scripting logic, and visual styling) that the computer-aided and automated detection of misuses of statistics and of misleading graphs can be further explored. State-of-the-art systems like LIDA evaluate their generated charts in six dimensions: code accuracy, data transformation, goal compliance, visualization type, data encoding, and aesthetics [1].