visit-dav / visit

VisIt - Visualization and Data Analysis for Mesh-based Scientific Data
https://visit.llnl.gov
BSD 3-Clause "New" or "Revised" License
436 stars 114 forks source link

Mark color tables that are good for color blind people. #17373

Closed brugger1 closed 2 years ago

brugger1 commented 2 years ago

Here is the e-mail thread for this request.

Eric,

Would this be something easy to implement in VisIt?

Thanks, Benjy


Hi Benjy,

Given the DE&I Moment in Ashley’s Bulletin on how our WebEx team added captioning capability to WebEx meetings, I thought I’d bring another seemingly-straightforward IT one to your attention.

We shows lots of simulation plots in WCI. Can the VisIt team determine and mark the palette options that also display well for people who are color-blind? The topic was discussed in this week’s APS mailing, see below. Since I’m not color-blind, I don’t know if there is more nuance to it than simply avoiding palettes that have what look to me like red-green combinations. I will say that some of the most commonly used palettes (e.g. “default” and “hot desaturated”) won’t work for people who are color-blind.

Best, Kumar


Hi Nick! I think this list is great. Thank you for taking the time to write it up and post this. I am wondering how we can get more interest in discussing these sort of topics in organizing plasma physics related meetings, like DPP. I am surprised no one else has dropped in. I think just the simple matter of being aware of what colors you use on plots/emphasizing use of patterns instead with colorblind people in mind is a simple step that can make a huge difference. I found this website recently to help me find appropriate RGB/HEX values and used this recently for a talk. Also, many computationalists are fans of "hot desaturated" as their colorbar of choice. Simply getting people to question their normal comforts and go-tos is a good and healthy thing. Another route to substantial behavior change is through course work. Say, for the off chance a professor is reading this: for course reports or projects perhaps have these "soft skills" (which are actually hard skills, apparently) in mind when developing your class projects and syllabi. I can see someone say, "but the physics!!!" but what does the physics matter if no one can understand what you're saying? I also believe that the ability to effectively communicate the physics illustrates true mastery. I can feel my eyes light up when I go to a talk and someone presents something in a creative and new way and I feel it click in my head, that's what I am always looking for at conferences. I am also curious... was IA15 able to implement these at following AAS Meetings? I am wondering how that went and how the AAS members felt about it, and whether members felt there was an noticable improvement, and if IA15 committee members were happy with the progress. It might be nice to have a discussion with DPP leadership and folks from IA15 if they're willing. Of course I am not a part of DPP leadership but I am happy to help to be a part of the solution!

Marissa Adams University of Rochester


Original Message: Sent: 02-04-2022 18:16 From: Nicholas Murphy Subject: Improving disability access at plasma conferences Following a few conversations after my last post, I wanted to expand upon how we can improve disability access at conferences. I wanted to share some of the practices that I've learned over the last few years on improving conference accessibility. I am hoping to spark a discussion that leads to accessibility policies becoming standard practice for plasma conferences, including for this fall's DPP meeting. More than that, I want accessibility and universal design to become core values in how we plan events and share information. I should emphasize that the list below is very much a work in progress. Accessibility policies need to be developed with input from both the disabled members of our community as well as subject matter experts in accessibility and universal design. Accessibility policies and practice need to evolve over time, in particular as assistive technology changes and various global pandemics happen. Also, an accessibility policy by itself is not enough. We need to change our community's culture, including removing the stigma associated with disability. • Hold the meeting in a space that at minimum meets ADA requirements. • Provide accessibility information on the conference website about the venue. Note any existing accessibility barriers. • Provide a straightforward mechanism for attendees to make requests for accommodations, such as ASL interpreters. • Have an accessibility point person or committee. • Make sure that the website meets accessibility standards, including for compatibility with screen reader (text-to-speech) technology. The WAVE tool provides automatic accessibility reports (such as this accessibility report from the current build of PlasmaPy's docs following some recent color contrast fixes). • Always use a microphone when one is available. Wait for the microphone to get to you before beginning to talk. Speak clearly into the microphone while facing the audience. Keep your lips visible for anyone who speech-reads. • Do not use flashing or strobing animations in presentations. • Tips for presentations o Use a large, easily readable font with sufficient color contrast. o Use dyslexic friendly sans serif fonts such as Helvetica, Verdana, and Arial. o Describe any graphics or figures out loud, and remark on important features. o Use colors that are accessible to anyone who is colorblind, in particular by avoiding red-green color pairings. For example, the hex values for a colorblind friendly triplet of colors for line plots is #920000, #006ddb, and #db6d00. o Provide captions for video content.
o Avoid using jargon and idioms. o Limit the information density on a slide, as too much information can be overwhelming for many. o Give sufficient time for conference participants to process the information. • Provide a quiet room that attendees can go to when needed. • Keep pathways clear for people who use wheelchairs or have limited mobility. • Minimize the use of fragrances and scented products. • Reserve the front row and aisle seats for people with accessibility needs. This could be done, for example, by marking those areas with blue tape. • Do not touch someone without permission and do not be offended if they refuse (even a handshake). This is especially important during the pandemic era. • Make sure that there are meal options for people with dietary restrictions. Include a place on the registration form to specify dietary restrictions. • Budget for accessibility needs, including potentially for closed captions, ASL interpreters, and transportation. • Make slides available to attendees ahead of presentations in an accessible format. • Make use of multimodal access. There are many situations where disability accommodations required by different participants are incompatible with each other. In these situations it is helpful to have multiple modes of access/communication/participation. In the pandemic era, this would translate to having options for both in-person and remote participation. A broader example would be making a book available both in printed form and as an audiobook. • Avoid planning conference excursions that involve activities that preclude the participation of disabled attendees. • Remember that not all disabilities are visible. A lot of these came from the accessibility policy for Inclusive Astronomy 2015, which has been adapted into the accessibility policies for a lot of astronomy meetings. I haven't yet compiled best practices for virtual meetings, however.

Nicholas Murphy Center for Astrophysics | Harvard & Smithsonian "Disability access is a human right. Lack of accessibility constitutes discrimination in and of itself."

markcmiller86 commented 2 years ago

Intersting questions regarding color tables.

Since we have folders for color tables, maybe we could create a folder with tables that work better for color-blind individuals? Do we know for sure which tables those would be. Are there metrics we can assign to them?

A question to circle back to Benjy...does WCI have any communication standards for this kind of thing? Does LLNL? I would think it would include things that go into the creation of PowerPoint slides or Posters or wherever our techinical work is communicated to a wide audience. Maybe someone in TID would have a handle on that. I am not trying to complicate things. Just understand where it fits in this larger context.

markcmiller86 commented 2 years ago

There are some pallete creation tools online, https://davidmathlogic.com/colorblind/#%23000000-%23E69F00-%2356B4E9-%23009E73-%23F0E442-%230072B2-%23D55E00-%23CC79A7

I worry there is a world of difference between pallet selection (e.g. discrete colors) for a UI or a set of curves in a 2D plot and a continous color mapping used in smooth data visualizations.

markcmiller86 commented 2 years ago

It looks like Ken Moreland has a nice article on line about scientific visualization

markcmiller86 commented 2 years ago

https://www.fabiocrameri.ch/colourmaps/

markcmiller86 commented 2 years ago

The above url references this download which contains many continuous and discrete color maps and has them defined for VisIt (*.ct format).

I saY we pull all these into VisIt into a "Vision Impaired" subfolder (I don't actually know if VisIt supports arbitrary color map folder names in the GUI. I see only Standard and User defined there now and maybe those are hard coded.).

We should document the color maps in our documentation and maybe provide a taxonomy for choosing color maps.

brugger1 commented 2 years ago

Here is some more information from an additional e-mail thread.

Thank Terri! I appreciate the info.... Contrast is definitely the big thing. I think many of the lab templates are fine but I really like your suggestion of "a document to staff about basic things they can do..." I find that it's the graphs and data displays prepared by individual presenters that are the most challenging for me!

Thanks again for the opportunity to give some input!

Steve


Greetings Steve,

I sent out some feelers to TID on Friday to gauge how our designers were handling this issue and if they received any specialized training in this area. This was the response I received:

This is a very misunderstood topic and TID is aware of it. We follow GSA 508 compliance guidelines for our documents. We haven’t received formal training, but we all know about the compliance guidelines. I always love learning, so I’d be interested in formal training if its available.

The most important thing we keep in mind while designing is contrast! Colorblind people can still see luminescence contrast so regardless of the colors we make sure the contrast is strong enough so they can see the difference even if its grey scale. Here is an article that explains it well. I suggest making sure your visual team is aware of the compliance and contrast issue if they aren’t already. Let me know if you have any other questions.

Benjy, should we confirm that the VisIT team is aware of the GSA guidelines? It’s quite in-depth. It also might be a good idea to provide a document to staff about basic things they can do to be inclusive in their design choices.

Looking forward to hearing everyone’s thoughts,


Hi Benjy,

Thanks for thinking of me. I'm happy to offer some suggestions.

As a little background, I suffer from two aspects of color blindness. I am both mostly red/green color blind but I also have an overall color perception deficit. The overall deficit means I have lower "cone" density and it's the cones that detect color.

What that all means is that I need to avoid red/green but it also means avoiding subtle shade differences. I can't see the difference between a light yellow and a light green without struggling. Same is true for a dark green and brown, there just isn't enough difference for me. If you're talking about lines in graphs, it's even harder because the lines are generally pretty thin...

So, basic rules for me are it's best to use primary colors and a lot of contrast. I can make red/green work, if they're hard primary colors and the lines are thick. A few good references and some good color pallet resources can be found here:

https://www.nature.com/articles/nmeth.1618. (see the table in Figure 2, those all work for me)

https://davidmathlogic.com/colorblind (a reference in the APS article) show IBM, Wong and Tol pallets (see bottom) and those are good too as long as the light level is high.

Using other optical features (like shapes, stripes, etc) can also help a great deal.

Hope that's helpful,

Steve

markcmiller86 commented 2 years ago

One thing to keep in mind is that color palette choices for UI are potentially very different from data display involving smooth shaded rendeering which may be different from photography/art.

All of those issues are likely useful to making VisIt more accessible to people with CVD but maybe different sources wind up informing different sets of choices.

markcmiller86 commented 2 years ago

From Francesca Semsel...

There is often a trade off between resolution and discriminatory power and whether an individual researcher is looking for a map that compensates for their specific type of colorblindness or one is trying to accommodate color blind folks. In general, CB folks can see value changes as well as blue-red or purple-yellow spans. The red blue provide more discriminatory power for non CB folks.

Most of the colormap testing considers only a few options. https://journals.plos.org/plosone/article?id=10.1371/journal.pone.0199239

I don't know of any work that has actually tested options in terms of which maps in category - linear, divergent, hue-cycling are most effective for colorblind folks. That would provide the best guidance.

I use this to check specific visualizations for colorblind folks: https://www.color-blindness.com/coblis-color-blindness-simulator/

A good tool for looking at colormaps is CCCTool. You can upload colormaps into CCCTool. Then use the CB Sim function to see how they translate. https://ccctool.com/html_v_0_9_0_3/CCC_Tool/cccTool.html

We ran tests on the colormap below. Interpolating using CIELab, it preformed as well as any others for colorblind folks while maintaining its discriminatory power for none colorblind viewers. That is the challenge.

image

markcmiller86 commented 2 years ago

Also from Francesca Semsel...

Colormap design has three disciplines from which to draw - the needs of the scientist, human perception principals and the one that folks always leave at the door… artistic color theory. Artistic color theory looks at the whole of the color interaction which in sci vis is really critical because the data, not a designer, controls which hues and in what quantity are presented adjacent to and in context with other hues.

A concern with the Crameri table suite is that it leaves out artistic color theory which dates back 600 years. The artifacts and misdirection of attention driven by the interaction of hues, a key principle, is not addressed

https://www.sculpting-vis.org

JustinPrivitera commented 2 years ago

completed in #17721