A Tool for Thinking in Systems
March 27, 2017 6:36 PM Subscribe
LOOPY (v. 1.0) is a new bowser-based tool from Nicky Case Previously, previouslier, more previously for constructing simple systems dynamics models by sketching; it also makes it easy to publish your work and let others modify it. Give it a try, or look at some examples: basic ecology; depression & anxiety; and automation & job loss.
"also it sort of works on tablets" Hah!
Very cool tool, which I can imagine using very profitably as a pedagogical tool in my near future. And a great sense of humor, too!
posted by dbx at 7:29 PM on March 27, 2017
Very cool tool, which I can imagine using very profitably as a pedagogical tool in my near future. And a great sense of humor, too!
posted by dbx at 7:29 PM on March 27, 2017
The automation example, completed: private killer bot army against plebeians, and government subsidy to mind-numbing circuses, shall do the trick, right? Max profits, max government, max joblessness, minimal unrest.
Yeah, my fixation on distopias managed to overcome myself, again.
posted by runcifex at 7:34 PM on March 27, 2017 [4 favorites]
Yeah, my fixation on distopias managed to overcome myself, again.
posted by runcifex at 7:34 PM on March 27, 2017 [4 favorites]
Cool! I had a job when I was a student making models like these in STELLA. it's neat that you can do modelling in a browser now, because I've wished I still had a copy of STELLA a couple of times since then.
posted by Dr. Twist at 7:34 PM on March 27, 2017 [4 favorites]
posted by Dr. Twist at 7:34 PM on March 27, 2017 [4 favorites]
This is very cool!
posted by turbid dahlia at 7:55 PM on March 27, 2017
posted by turbid dahlia at 7:55 PM on March 27, 2017
I really wish it had a few extra math functions like multipliers and maybe a way to distinguish sources and you could actually do PID control and the like. I mean you could with what you have kinda but that would be some pain in the ass binary stuff.
posted by Puddle at 8:28 PM on March 27, 2017 [2 favorites]
posted by Puddle at 8:28 PM on March 27, 2017 [2 favorites]
Yeah, this is a really cool project with a pretty minimal feature set. Someone shared this with me last week while I was working on a presentation, and I immediately tried to make some graph-gifs with it, but was stymied by the inability to distinguish sources as well as the way the directions of the arrows coming out of a node are determined by the directions of arrows coming in.
I hope they continue to add features, because it's just excellent otherwise.
posted by galaxy rise at 9:13 PM on March 27, 2017
I hope they continue to add features, because it's just excellent otherwise.
posted by galaxy rise at 9:13 PM on March 27, 2017
Yeah, my fixation on distopias managed to overcome myself, again.
Increased tax revenue from greater corporate profits is offset by reduction of tax revenue from job losses. The neutered government is unable to provide for the commonweal. Dystopia.
posted by um at 1:16 AM on March 28, 2017
Increased tax revenue from greater corporate profits is offset by reduction of tax revenue from job losses. The neutered government is unable to provide for the commonweal. Dystopia.
posted by um at 1:16 AM on March 28, 2017
The people of Dystopia elect a Social Democratic style government that increases the corporate tax rate: Fully Automated Luxury Communism.
posted by um at 1:24 AM on March 28, 2017 [3 favorites]
posted by um at 1:24 AM on March 28, 2017 [3 favorites]
It seems that when a thing reaches zero it loops round to being full again. Is this expected?
Can I make it not do that?
posted by Just this guy, y'know at 3:54 AM on March 28, 2017
Can I make it not do that?
posted by Just this guy, y'know at 3:54 AM on March 28, 2017
the directions of the arrows coming out of a node are determined by the directions of arrows coming in.
Yeah, that was not what I was expecting at first either, but it does make sense now that I've played with it further.
I built this thing to get a better idea of how it interacts.
The inverter changes each cycle from a positive cycle to a negative one, so first time round the positive adds and the negative subtracts, and then next time round it swaps.
Hmm, actually I guess it would be better to see if each chunk only did one action.
Like this!
posted by Just this guy, y'know at 4:21 AM on March 28, 2017
Yeah, that was not what I was expecting at first either, but it does make sense now that I've played with it further.
I built this thing to get a better idea of how it interacts.
The inverter changes each cycle from a positive cycle to a negative one, so first time round the positive adds and the negative subtracts, and then next time round it swaps.
Hmm, actually I guess it would be better to see if each chunk only did one action.
Like this!
posted by Just this guy, y'know at 4:21 AM on March 28, 2017
I never really got Systems Dynamics - seems (like everything else) very very dependent on the premises. Like 'rhetoric plus'. How does one argue the validity of these models?
posted by anthill at 4:31 AM on March 28, 2017 [2 favorites]
posted by anthill at 4:31 AM on March 28, 2017 [2 favorites]
How does one argue the validity of these models?
I was pondering this line too, but it seems like the value is really in the relationship visualization.
Ie, if you have a bunch of pieces and want to have a more cohesive view of how they fit together. The individual links all have their own empirical questions. But as a whole, if you can take some things as tacit and allow yourself to make some things 'fixed', then you can use this to draw a bigger picture of your pieces.
@Systems folks, is that kind of it?
posted by Reasonably Everything Happens at 5:30 AM on March 28, 2017 [1 favorite]
I was pondering this line too, but it seems like the value is really in the relationship visualization.
Ie, if you have a bunch of pieces and want to have a more cohesive view of how they fit together. The individual links all have their own empirical questions. But as a whole, if you can take some things as tacit and allow yourself to make some things 'fixed', then you can use this to draw a bigger picture of your pieces.
@Systems folks, is that kind of it?
posted by Reasonably Everything Happens at 5:30 AM on March 28, 2017 [1 favorite]
MiFi Trump Thread Simulation.
(best if clicked up to high speed ;-)
posted by sammyo at 5:58 AM on March 28, 2017 [3 favorites]
(best if clicked up to high speed ;-)
posted by sammyo at 5:58 AM on March 28, 2017 [3 favorites]
What I like about this is clicking the circles makes arrows come out I lover clicker games.
posted by Tevin at 6:04 AM on March 28, 2017 [2 favorites]
posted by Tevin at 6:04 AM on March 28, 2017 [2 favorites]
Systems folk here.
You do want each individual arrow to have some validation behind it, whether it's expert opinion or clear correlations in data. The order I was taught was that you begin with a question about behavior over time ("Why does surging during a crisis cause the crisis to get worse over time?"), then build as small a model as possible that captures the relationships that affect the behavior of interest. Validate/justify the assumptions that go into the model with the people who want an answer, and they'll generally trust the insights that come from it ("Because nobody ever gets credit for fixing problems that never happen", which happens to be a very readable paper about one application of System Dynamics). One nice thing about these models is that they're easy to adjust if you think there's a violated premise: we can see what behavior each set of assumptions exhibits, compare it to lived experience, and go with the more predictive option.
posted by persona at 10:55 AM on March 28, 2017 [5 favorites]
You do want each individual arrow to have some validation behind it, whether it's expert opinion or clear correlations in data. The order I was taught was that you begin with a question about behavior over time ("Why does surging during a crisis cause the crisis to get worse over time?"), then build as small a model as possible that captures the relationships that affect the behavior of interest. Validate/justify the assumptions that go into the model with the people who want an answer, and they'll generally trust the insights that come from it ("Because nobody ever gets credit for fixing problems that never happen", which happens to be a very readable paper about one application of System Dynamics). One nice thing about these models is that they're easy to adjust if you think there's a violated premise: we can see what behavior each set of assumptions exhibits, compare it to lived experience, and go with the more predictive option.
posted by persona at 10:55 AM on March 28, 2017 [5 favorites]
There is also a very good blog about the why's and wherefores of building it.
posted by Just this guy, y'know at 11:00 AM on March 28, 2017 [4 favorites]
posted by Just this guy, y'know at 11:00 AM on March 28, 2017 [4 favorites]
This is so great! Thank you.
posted by the man of twists and turns at 8:45 PM on March 28, 2017
posted by the man of twists and turns at 8:45 PM on March 28, 2017
« Older Adapting ASL for music | Shiny Happy Robots Holding Hands Newer »
This thread has been archived and is closed to new comments
Woof?
posted by hank at 7:05 PM on March 27, 2017 [3 favorites]