The way the final dashboards (the end result) are done uses strategies that, as 2020, are very size-optimized compared to other alternatives for serve data on web. The strategies (the design chooses) already are very optimized, even for an alpha release.
This issue actually is the part that are not the end result, but the screens that are related to explain how to use the dashboards (and also about the home page). At this point of discussion is just about the images.
Strategies to optimize images
There are some ways to optimize images. Some can be integrated at server size (like pagespeed modules), other can be added at building times (like some commands with nodejs or other tools.
One way that may work on hxldash
As for the hxldash, since the number of images is small (and is not user generated) maybe from time to time one way could be just upload the images for some services (like https://tinypng.com/) check if the % of images worth the optimization, and then
The next images are not really accessible to screen readers, but a description of what they show is a list of images (and original sizes) vs size of new images and the saving on the file size after the optimization. The smallest savind is 11%, and the biggest is 88% (not sure the averagen size reduction, but may be at least 33% or more for the #67 pull request.
The way the final dashboards (the end result) are done uses strategies that, as 2020, are very size-optimized compared to other alternatives for serve data on web. The strategies (the design chooses) already are very optimized, even for an alpha release.
This issue actually is the part that are not the end result, but the screens that are related to explain how to use the dashboards (and also about the home page). At this point of discussion is just about the images.
Strategies to optimize images
There are some ways to optimize images. Some can be integrated at server size (like pagespeed modules), other can be added at building times (like some commands with nodejs or other tools.
One way that may work on hxldash
As for the hxldash, since the number of images is small (and is not user generated) maybe from time to time one way could be just upload the images for some services (like https://tinypng.com/) check if the % of images worth the optimization, and then
My initial suggestion
See Pull Request #67. At this point, I just used the https://tinypng.com/.
The next images are not really accessible to screen readers, but a description of what they show is a list of images (and original sizes) vs size of new images and the saving on the file size after the optimization. The smallest savind is 11%, and the biggest is 88% (not sure the averagen size reduction, but may be at least 33% or more for the #67 pull request.