RE: ARIA roles for charts (was: Master thesis on SVG accessibility)

I think this is some great stuff. Happy to hop on a call about this with you, but I think it's important to work backwards from functional criteria/goals, IMHO.

If your only goal is textual access, that's not sufficient, at least to me, to be honest, as simply providing the source data is basically that, and textual access is nice for examples and demos, but impractical in the real world if one needs to view ten different datasets, let's say for stock prices, and then quickly come to some conclusions. Apple has made some wonderful strides on this with their automatic data comprehension features in iOS 13. It offers a numerical summary, a description of the data, some help text to explain what's going on, and a data sonification, all one rotor movement and a flick away, once you land on chart-data. don't get me wrong, textual access is critical, but it's just not a way I feel you could make the most impact with your research. It's necessary but not sufficient, to use a common phrase in the field.

So, I submit that a functional goal should be the ability to understand, or comprehend, the data, not only iteratively, but in terms of the gestalt of the info. The math to do this is of course trivial .. basic statistical analysis is essentially free given a variety of different libraries. Sonification has also been solved half a dozen times repeatedly, so I would urge you to start from some existing progress in the field on that front as well. Happy to make any necessary introductions if you don't already have those contacts. Let's move things forward, not reinvent wheels that have been constructed in slightly different ways since the late 90s, possibly sooner 😊.

In an ideal world, we'd have an HTML concept of charts. I know ChartML is a thing, but that's not HTML obviously and it's not used often on popular websites, unless if I'm mistaken, which if true would delight me, so someone please feel free to yell at me if I'm wrong about that. Barring the html concept, something HTML-friendly at least, would go a long way, which of course brings us to your SVG ideas building off of Doug's work. The use case I imagine involves going to a page, encountering a chart, and having it behave much like MathML solutions do, which is to say an interactable region one can enter/exit out of, but within which there exists a rich set of affordances. If one wishes a simple textual view, sure, that's easy, but if one wishes to hit s, for sonification, and hear the graph or move along and explore the data, etc., then that's delightful. If one is low vision and wishes to only zoom in the graph, that can obviously be done easily, and if one is wanting a textual summary of the graph, which would benefit overview purposes for blind users as well as explanatory purposes to help with cognitive differences amongst users, then this can easily be requested/generated/viewed as well. An HTML representation such as a table or grid could also easily be provided as a trivial runtime transformation, which would help in exploring the dataset, especially for screen reader users (just don't confuse exploration with comprehension). Additionally, if this work is combined with some of the exciting stuff around braille version of role/label information, then one can have a concise braille interpretation too. This is made even more exciting with the continuous rise of full-page Braille displays. Also, if you can do focus highlighting when exploring the data, then this has the potential to greatly help with graph literacy for various learning differences, but you'd need to validate any such claims with evidence, of course. Ph.D. in your future? *smile*

Note , some of the above is done currently by SAS's Graphics Accelerator Chrome extension. It would be great to have a fully free and open source alternative to that pipeline.

The above only scratches the surface. For this to be actually useful, in my opinion, it would need to then be integrated into C3/D3 libraries, etc., and then you could make a real, actionable, honest to goodness, dent in the inaccessibility of 99+% of data visualizations out there.

So, now to discuss roles. Do you have an enumeration of the various data visualizations you wish for the proposed set of roles to work for? Concepts such as points, bars, axis, start/end positions, areas in graphs (imagine a shaded region), percentages of the whole (totally seems like posinset  could be extended to deal with percentages), summaries, representation concepts like zoom, and maybe steppings for various iterative navigation all come to mind, but that's hardly exhaustive.

I imagine some research or lit review has been done so you know the top 5 used concepts in this space, or something like that, to help aim and constrain your efforts, since the goal should be to graduate, not boil the ocean, *smile*. Sorry, I'm AbD in a Ph.D. in HCI, so I feel this need to get grad students graduated. No patronization intended.

Since you're a masters student, if you'll permit me to suggest to you some reading in this space, I would make sure you are abundantly familiar with the great work that has come out of Bruce Walker's lab at Georgia Tech. Just go read the last 10 years of papers there. From point estimation to using abbreviated phonemes for greater speech fluidity/efficiency to all sorts of sonification-related explorations, there's good stuff there. Also make sure the ICAD conference is on your radar if sonification is at all of interest. If not, then just treat it as a black box, but include it, even if you don't wish to innovate in that space, which is totally reasonable given that you wish to concentrate on these other aspects. There's tons of other papers that could be of use (the Assets conference of course), but it would be good to know where you wish to concentrate, and presumably where your prof wishes you to concentrate, before I suggest more.


This is a high-level response, and I know you specifically asked about roles, which I would be happy to geek out about, but I just want to make sure I understand the exact scope of the problem you wish to solve, before diving into implementation land. Hopefully some of the above is useful or sparks something for future conversation. Keep in mind that no matter what you do, you are mapping a high-dimensional space onto a low dimensional space, so anything you can do with multimodal output, flexibility of interaction, query-based mechanisms, and the like will be a step towards equalizing that bandwidth gap.

Oh, also, if this conversation is inappropriate for this list, then my apologies to the mods, and obviously, I'd be happy to move it off list, but you have some of the smartest and hardworking accessibility folks on this mailing list, so it sure seems like a nice starting place for the discussion, in my opinion.

Thank you for wanting to make the web and data visualizations on it more accessible. That's a meaningful contribution no matter how you choose to pursue this project, and you have my sincere appreciation.

Take care,
Sina

President, Prime Access Consulting, Inc.
Phone: 919-345-3832
https://www.PAC.bz
Twitter: @SinaBahram
Personal Website: https://www.sinabahram.com

-----Original Message-----
From: Christopher-A. Kopel <kopel@student.tugraz.at> 
Sent: Monday, November 4, 2019 11:43 AM
To: public-svga11y@w3.org
Subject: ARIA roles for charts (was: Master thesis on SVG accessibility)

Hi all!

I'm picking up the thread I started in September 2018 about a master's 
thesis on accessible SVG (see my initial post below).

First, thank you for all your replies!

I'm currently working hard on the practical part of my thesis. This 
shall be a web-based tool (and in the future maybe a browser extension) 
which parses charts with ARIA markup and presents them in a textual way 
with (hopefully intuitive and easy-to-use) keyboard navigation. The UI 
shall be usable both in combination with a screen reader and stand-alone 
with speech synthesis provided by the Web Speech API. The tool is based 
on Doug Schepers' "Describler" [2].

Now one essential question is which system of ARIA roles to assume for 
recognising charts and their child elements such as data series, axes, 
and data points. It seems to me that there are several approaches 
towards this but not yet any standard or at least any trend towards it. 
Here's what I found:

- the proposed system "ARIA Roles for Charts" - still only a Wiki since 
2015 [1]
- Doug Schepers' system of ARIA roles and properties used for Describler [2]
- Doug Schepers' new charting suite "fizz.studio" using the standard 
roles "graphics-document", "graphics-object", and "graphics-symbol", 
refining them with various values for "aria-roledescription" [3] (no 
official taxonomy or sample charts yet)
- a proposal by Léonie Watson to annotate chart elements with 
corresponding ARIA roles for tables, such as "table" for the chart, 
"row" for a data series, and "cell" for a data point [4]
- a system applied by Highcharts using role "region" for data series and 
role "img" for data points but no roles for axes, legends, or the chart 
itself [5]

Of course, it might be best if my tool could cope with all these 
different types of chart encoding formats one day, but now I'm wondering 
which one to start with. Do you have any suggestion?

What about the proposed system presented in the Wiki [1]? - Is it still 
meant to become a standard, or has it been dismissed in the meantime?

If you know a better place to ask this (maybe the public-aria list), 
please tell me!

Many thanks in advance for your help and your ideas!

Cheers,

Chris


[1] https://www.w3.org/wiki/SVG_Accessibility/ARIA_roles_for_charts
[2] http://describler.com
[3] https://fizz.studio/
[4] https://tink.uk/accessible-svg-line-graphs/
[5] https://www.highcharts.com/a11y.html


-------- Weitergeleitete Nachricht --------
Betreff: Master thesis on SVG accessibility
Datum: Wed, 26 Sep 2018 00:59:21 +0200
Von: Christopher-A. Kopel <kopel@student.tugraz.at>
An: public-svga11y@w3.org

Hi all!

I am student of Information and Computer Engineering at the Graz 
University of Technology (Austria) and intend to start my master thesis 
in October. As I am blind, I would be highly interested in writing about 
a topic in the field of web accessibility.

One of my professors, who mainly researches on data visualisation, had 
the idea that my thesis could deal with the accessibility of SVG, which 
seems very interesting to me, too. However, reading some publications 
about this topic (i.e., some of the recent W3C recommendations or 
recommendation proposals, respectively, as well as some texts and 
presentations by Doug Schepers and Léonie Watson), I got the impression 
that much good work has already been done on this. For this reason, I'd 
like to ask you for your opinions: Do you think that with a master 
thesis on the accessibility of SVG I could contribute to its 
improvement? If so, is there any work for me which you would regard as 
particularly valuable at the moment? It would be great if this could 
include any kind of implementation but I'm also fine with some analysis, 
testing, or evaluation. For example, I think of a tool (or maybe even 
better, a patch to a browser or screen reader) that transforms all 
usable information of an SVG document into a form that is 
well-comprehensible for a blind user, taking Doug Schepers' "Describler" 
as a starting point. Another idea of mine would be a patch to an SVG 
authoring tool or an SVG scripting library like 3d.js that improves the 
accessibility of the resulting SVG structures.

Thank you very much in advance for your feedback and any suggestions!

Cheers,

    Chris

-- 

  *** Message from: ***

Christopher-Alexander Kopel, BSc
E-Mail kopel@student.TUGraz.at

Received on Monday, 4 November 2019 18:33:27 UTC