Optical Illusion?

[cheddar]

I was scrolling through some data to verify that things are working properly and noticed the recommended basals for the first time (can we say "recommended" instead of "recom'd"?), yay.

The one I was looking at made me wonder why recommended was being shown as light blue inside of a dark blue border, when the legend made me think that I should see two heights, one for the actual and one for the recommended. Then I got to thinking and wondered if the light blue was maybe going up to the height of the recommended and the dark blue was going up to the height of the actual. I started scrolling around to look at more data points and try to verify/disprove this, and I believe I verified it.

But, I believe it also elucidated the actual head-tilting thing. Which is why I am writing this. In the screenshot above, there are two boluses that ignore the recommendation toward the right. To your eye, what is the ratio of the right-most bolus's actual delivered amount to the recommended amount from the bolus to its left?

To my eye, the right-most bolus actually delivered maybe 1/2 to 2/3 of the recommendation from the bolus to its left.

When I hovered over the tooltips, this was not corroborated though, which was interesting. So, I got out a straightedge and sure enough, it's actually drawn to the correct scale. The ratio is that the bolus on the left recommended 3.8 units and the one on the right actually delivered 3.6 units, so they are actually equivalent in height. This gots me believing that we have ourselves an optical illusion going on here. Do others see it too?

[jebeck]

I don't think I suffer from the optical illusion, but I'm probably not the best judge, so...I'll leave that aside.

The other issue these comments raise for me, however, is whether we want to add some additional explanation of how underrides and overrides are represented (I believe @brandonarbiter @skrugman we discussed this once?). This could be done in a tooltip when hovering on the legend, for example.

[kentquirk]

I'm not sure it is an optical illusion.

If you zoom in:

The left box is 16 pixels high. The right one is 9 pixels high.

Does this make sense based on the data?

On Fri, May 2, 2014 at 8:32 PM, Jana Beck notifications@github.com wrote:

I don't think I suffer from the optical illusion, but I'm probably not the best judge, so...I'll leave that aside.

The other issue these comments raise for me, however, is whether we want to add some additional explanation of how underrides and overrides are represented (I believe @brandonarbiter https://github.com/brandonarbiter @skrugman https://github.com/skrugman we discussed this once?). This could be done in a tooltip when hovering on the legend, for example.

— Reply to this email directly or view it on GitHubhttps://github.com/tidepool-org/hub/issues/73#issuecomment-42091403 .

Kent Quirk VP of Engineering, Tidepool

Tidepool is an open source, not-for-profit effort to build an open data platform and better applications to reduce the burden of Type 1 Diabetes.

[brandonarbiter]

I think the tooltip makes sense as is. I might suggest one addition. Currently, the tooltip says

XU YU recom'd

Where "XU" is the amount of insulin delivered and "YU" is the amount of insulin recommended. For "YU", we clarify using the word "recom'd". However, there is no such clarification for "XU". Perhaps we can try looking at a design where we add the word "dlv'd"? @jebeck @skrugman

@jebeck Can you take a look at the pixelation @kentquirk points out and verify that ratio is correct?

On Fri, May 2, 2014 at 6:15 PM, Kent Quirk notifications@github.com wrote:

I'm not sure it is an optical illusion.

If you zoom in:

[image: Inline image 1]

The left box is 16 pixels high. The right one is 9 pixels high.

Does this make sense based on the data?

On Fri, May 2, 2014 at 8:32 PM, Jana Beck notifications@github.com wrote:

I don't think I suffer from the optical illusion, but I'm probably not the best judge, so...I'll leave that aside.

The other issue these comments raise for me, however, is whether we want to add some additional explanation of how underrides and overrides are represented (I believe @brandonarbiter < https://github.com/brandonarbiter> @skrugman https://github.com/skrugman we discussed this once?). This could be done in a tooltip when hovering on the legend, for example.

— Reply to this email directly or view it on GitHub< https://github.com/tidepool-org/hub/issues/73#issuecomment-42091403> .

Kent Quirk VP of Engineering, Tidepool

Tidepool is an open source, not-for-profit effort to build an open data platform and better applications to reduce the burden of Type 1 Diabetes.

— Reply to this email directly or view it on GitHubhttps://github.com/tidepool-org/hub/issues/73#issuecomment-42092529 .

[jebeck]

I can't really do any verification without knowing the source of this data (whose and what date)... @cheddar do you recall?

[cheddar]

Hrm, actually, I think I did a poor job of explaining my illusion. I believe @kentquirk counted the full dark blue line (the actual delivered) on the left side. I was referring to the light blue (recommended) on the left side.

I.e. the size comparison was supposed to be

left: light blue right: dark blue

I just verified with Kent over irc that the light blue on the left is 10 pixels, which is fine given that it's just slightly more than the amount delivered for the dark blue on the right.

[jebeck]

So are there any remaining action items on this issue, or could it be closed?

[cheddar]

Can be closed. Seems like it's just me.