19 Jun 2019 |
martin | Download Screenshot from 2019-06-19 16-17-16.png | 06:22:06 |
martin | Download Screenshot from 2019-06-19 16-22-30.png | 06:22:52 |
martin | nodes can also be dragged to arbitrary positions, and you can use ctrl/cmd Z ctrl/cmd Y to undo and redo actions | 06:30:48 |
Matt Shin | Undo and redo! That's great. | 06:45:56 |
Matt Shin | Now that I've seen more of it, I'm less keen on the very colourful scheme. It is quite distracting, to be honest. I now prefer a more monochrome scheme that uses light to dark shades to represent progress. | 06:49:43 |
Matt Shin | Otherwise, great progress! 👍🏼 | 06:53:12 |
@kinow:matrix.org | 👍 looking great! | 07:00:00 |
Hilary Oliver | Yeah, great progress martin ; thanks. | 07:05:26 |
martin | the colour scheme is based on oliver sanders lovely diagrams 🙂 | 23:48:08 |
20 Jun 2019 |
Hilary Oliver | Colours look alright to me. And are easy to change anyway! | 05:33:10 |
martin | Redacted or Malformed Event | 05:34:21 |
martin | Download simple-cytoscape-dot.4.json | 05:35:02 |
martin | oooh thaty's ugly - here's an upload | 05:35:21 |
martin | the JSON data used at present for the cytoscape graph | 05:35:39 |
Matt Shin | https://www.poynter.org/archive/2013/why-rainbow-colors-arent-always-the-best-options-for-data-visualizations/ https://agilescientific.com/blog/2017/12/14/no-more-rainbows
Just some readings on why rainbow's not good for data visualization. | 05:57:32 |
Sadie Bartholomew | Interesting reads! I like that last link's title. Sounds like a perfect Twitter hashtag: #nomorerainbows (as long as the context was clear from the tweets). | 09:31:42 |
Sadie Bartholomew | Something I learnt about last Hacktoberfest is that often colourmaps are based on the Red-Green-Blue space, but that there are alternatives to RGB which re-map to define colours along three other axes. The one that sems most promising is HCL: if interested, see e.g. http://hclwizard.org/why-hcl/ (admittedly a bit biased) & the paper (actually of meteorological context!) where some of those visuals & schematics are from: https://www.researchgate.net/publication/275467530_Somewhere_Over_the_Rainbow_How_to_Make_Effective_Use_of_Colors_in_Meteorological_Visualizations | 09:37:24 |
Sadie Bartholomew | "Somewhere Over the Rainbow: How to Make Effective Use of Colors in Meteorological Visualizations"... so I guess it follows on nicely from those #nomorerainbows reads. | 09:38:53 |
Tim Pillinger | Viridis is now standard in Scipy: It might be interesting to look at the posters around the office with my colour-blindness app comparing viridis to jet. | 09:54:06 |
Sadie Bartholomew | Ultimately, for our GUIs, (I believe) we are looking for a discrete (& very small, ideally) set of colours, not a colour map, though. We're not doing visualisations as such; we just want a consistent scheme. So really it is about a set of colours which work well together & can't be distinguished by those with accessibility issues. It would not have to be pulled from a colour map. | 09:59:26 |
Sadie Bartholomew |
| 10:12:02 |
oliver sanders |
Just some readings on why rainbow's not good for data visualization.
| 10:48:26 |
oliver sanders | True, we aren't actually using colour series so for us it's not about data visualisation but visual association | 10:49:04 |
21 Jun 2019 |
martin | https://color.adobe.com/create | 01:35:28 |
Hilary Oliver | Yes "color maps" aren't very relevant to us. And we're looking a fewer colors too, this time, not one for each of the 11 task states. | 04:54:36 |
Matt Shin | Yes and no. It is relevant for accessibility and for clarity. We should probably only have two sets of palettes. One set of colour scales for the normal progress states and one set of colour scales for the error conditions. | 05:42:35 |
martin | Download Screenshot from 2019-06-21 16-46-46.png | 06:49:24 |
martin | Download Screenshot from 2019-06-21 16-45-17.png | 06:49:25 |
martin | Download Screenshot from 2019-06-21 16-45-50.png | 06:49:26 |
martin | Download Screenshot from 2019-06-21 16-45-00.png | 06:49:26 |