-
Notifications
You must be signed in to change notification settings - Fork 25
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
1316 observed area #2860
1316 observed area #2860
Conversation
…into 1316-observed-area
…into 1316-observed-area
…into 1316-observed-area
…into 1316-observed-area
@jonfroehlich @misaugstad Thoughts? |
Yay! So great to see this revived! I think this is really thoughtful and I'd like to try it out to get a better sense of things:
Sometimes, however, it is useful to see things from two different perspectives (e.g., looking at something from the south-to-north then taking a step north and looking back north-to-south)—I find this particularly helpful for surface problems. |
@jonfroehlich That makes sense—didn't think of that! |
I don't expect us to prevent people from moving before they've spun around (this would be incredibly frustrating on streets with no sidewalks at all). So I agree that it's a bit harsh ;) I think that what's most important to us right now is that it's shown in a visually clear and appealing way, and that we are able to record this information so that we can use it later. We might start by just recording a percentage for each pano or possibly street. Then once that's live for awhile, we might then explore with that data a bit to see if it has a predictive effect on data quality, and come up with exactly we want our metric to be at that time. And we would likely add this to a user's dashboard. Possibly the leaderboard. And possibly as a part of our heuristics for determining data quality from a user.
I have to agree with @jonfroehlich on this. If one sidewalk block is raised above an adjacent one, that's generally really clear when looking at it from the side with the lower block, and often very hard to see from the side with the raised sidewalk block. As such, doing a strict "observed area" method like you have right now would have the downside of discouraging users from looking backwards. I just tried out your PR. It looks so good!! If turning around to check from another angle wasn't so important, I would be 100% all in on the current method. I'm trying to think about possible ways to visualize that you've looked through your previous panos successfully... One idea could be to a small, circular "progress bar" thing at each panorama location that you enter. And those could persist on the mini-map after you leave the pano. To encourage you to leave a trail of fully complete progress bars in your wake. But that adds a decent amount of visual clutter. Maybe the fog of war could have 3 levels instead of 2? One for not being explore, one for when it has been looked at from one angle, and another for when it's been looked at from both directions? Not sure how complicated that is to code up... I guess it would probably look like having two objects holding "observed areas"; one for areas you observed that were in front of you, and one for observed areas that were behind you. And you could use the angle you should be facing to move forward down the street to determine in front / behind. Though this could get wonky at intersections 🙃 Idk just spitballing here. Any thoughts @shokiami ? |
@misaugstad I'm glad you like it :) and that makes sense! I just wanted to clarify our intended use of the observed area thoroughness data before making further assumptions about the problem. So I'll go ahead and switch it back to using angles (unless there is some other metric you want me to use). |
Re: cycles. I meant because I thought your primary focus was on CV and also
that you have an internship coming up soon. :)
On Thu, Apr 14, 2022 at 8:59 PM Shosuke Kiami ***@***.***> wrote:
@misaugstad <https://github.com/misaugstad> I'm glad you like it :) and
that makes sense! I just wanted to clarify our intended use of the observed
area thoroughness data before making further assumptions about the problem.
So I'll go ahead and switch it back to using angles (unless there is some
other metric you want me to use).
I agree that adding a circular progress bar for every previously observed
area would be a lot of visual clutter. I like what Jon is suggesting though
which is to add a single progress bar for the current location. If we were
to do this, do you think it would look better in the sidebar or the
mini-map?
Ooh, I'm intrigued by this idea of having 3 levels of the fog-of-war! It
might get complicated, but I'll do my best to implement it to see how it
looks. Also, do you think that this should have any effect on our observed
area thoroughness data we are collecting?
@jonfroehlich <https://github.com/jonfroehlich> I'm more than happy to
continue working on this project! A little confused about what you mean by
"cycles"... I'm still in my second year if that answers your question.
—
Reply to this email directly, view it on GitHub
<#2860 (comment)>,
or unsubscribe
<https://github.com/notifications/unsubscribe-auth/AAML55LHGUNGT6MCHIVIZHLVFDSQDANCNFSM5TJFJVRA>
.
You are receiving this because you were mentioned.Message ID:
***@***.***>
--
Sent from phone
|
…into 1316-observed-area
@jonfroehlich @misaugstad Here is what everything looks like right now: To summarize, this PR introduces:
Note: this PR also resolves #2329 |
Sounds good to me! Kudos for continuing to push this forward. Can you create a new Issue for discussing how to incentivize people to look backwards at an area after taking "a step" forward? |
@jonfroehlich Just did: #2982 |
I'll need to do some testing and review the code. @shokiami do you want me to do the full PR review? Or if there are just some code organization/efficiency changes to make, should I just make them myself? Not sure what your internship workload is looking like. |
@misaugstad Feel free to do the full PR review! |
If that doesn't work, you might have to draw a black outline around the lettering and circle. I can help with that if relevant in the future. |
After a quick look, @shokiami and I talked about this and we agreed that the code for the mini map should live in it's own file, and we can put it in the /navigation directory. @shokiami if you want to start moving some of the other functionality related to the mini map into that file, feel free, but that can also be separated out into it's own issue later on! |
…into 1316-observed-area
…alkWebpage into 1316-observed-area
@misaugstad @jonfroehlich |
Very cool @shokiami. I can't wait to ship this long-awaited feature! 🚀 One thing I just thought of--which would be simple to add just not certain on the text, which would need to be very concise and clear: |
@jonfroehlich |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Awesome work, @shokiami! This looked really good. The restructure was great. We were so close to being done, and I wanted to push out a new version of Project Sidewalk this week, so I did the final little edits myself so that we could get it pushed out!
Some changes:
- All the mini map stuff was called
google-maps
orgoogleMap
, which felt confusing, so I renamed all of those tominimap
. That's part of your code specifically, but it's an extra little change I made on top of it. - I switched up how you keep track of the list of observed areas. Instead of making sure that the current one is at the end of the list, I just made a new
currArea
variable that keeps track of that. And we now use thepanoId
as the index instead of thelatLng
. - I didn't like how the green fog of war looked, it was a pretty jarring change. I switched it back to having a gray fog of war always, and making the circle turn green when you reach 100%. I just kept the text black, since the main issue with the green progress circle was that you couldn't read the text when it turned green.
Awesome, thank you so much @misaugstad! |
Resolves #1316
Resolves #2329
After a long hiatus (sorry for the wait 😅), I finally implemented a rough draft for the observed area PR!
Here's what it looks like:
One design decision I made that deviates from what we have talked about is measuring the observed area percentage in terms of the area instead of angles. For example, the image above depicts a user who just took a step while auditing, and although the user has not yet swiveled around in their new position, since the observed area from the previous position bleeds into the new one, it measures their current observed area as being ~83%.
Also, here are some relevant feeds:
#1316
#1324
#2329