search

CityScope / CS_Cooper-Hewitt

meta repo/sandbox repo for keeping everything related to the Cooper Hewitt exhibition
5 stars 2 forks source link

Congestion visualization #105

Closed yasushisakai closed 5 years ago

yasushisakai commented 5 years ago

Gives you this:

c7eea379-8b4f-4579-b11f-5eac36c7686b

agrignard commented 5 years ago

@yasushisakai it sounds god and a lot fo modification in the code. How does it behave in term of performance?

I would make the same remark as I did for this PR https://github.com/CityScope/CS_Cooper-Hewitt/pull/104. Let's first focus on having a release of the plan A today and then we can go to A+ and even more by pushing some optimization and new rendering (but it needs to be validated by everyone first) but not before having a final release A

As I said from the beginning if we have to surimpose a heatmap on the agent that means we have miss something from the beginning in the design , if you think about a heatmap it is supposed to be an abstraction/aggregation of something and in our case we show the abstraction (the heatmap) and this something (the agent). I am afraid ti will be to much and or redundant (I mean we can already see that there is lot of car no?) but we have to test on the table for sure, by running this branche.

So I wouldn't merge this one either to be sure not to introduce last minute debugging (we cannot afford this anymore)

agrignard commented 5 years ago

Just by testing it quickly I see a regression in the rendering of the network (just press 'n') and you will see.

screen shot 2018-11-26 at 07 20 01

What I propose is to iterate in this branch but it needs more test to be sure it's stable . Maybe we have to think of having different computer running different branches to validate them.

Let's talk this about this today (for the design it needs to be discuss) I like the programmatic solution but I am not 100% sure of the final rendering

yasushisakai commented 5 years ago

Yes, indeed, this needs careful scrutiny

yasushisakai commented 5 years ago

Not to impose anything, but 469d929 fixes the thick ‘n’ show network function.

agrignard commented 5 years ago

Cool 'n' is working now.

Sometime some black rectangle are displayed and then we don't see anymore the background I don't think we want this, maybe not display the rectangle if black?

screen shot 2018-11-26 at 11 45 06
yasushisakai commented 5 years ago

I’ll take a look in the morning, my mouth hurts. (took a tooth) 🔧 👄 🔨

yasushisakai commented 5 years ago

Something has changed to the master 469d929 can no longer be applied...

I tried to make another pull request branching from 469d929 and merge it with the current master, but as I solve the conflicts, agents will not move anymore.