Segregation

[Alan Isaac]

Use this thread to discuss the Schelling segregation model and the next assignment (which is based on it).

[Alan Isaac]

Now that you are starting to work with BehaviorSpace, be sure to reread
https://subversion.american.edu/aisaac/notes/netlogo-intro.xhtml#behaviorspace-cautions

Please be sure to ask questions about your use of BehaviorSpace as soon as possible.
Do not wait until the last minute!

[Adaora Isichei]

Hey Professor Issac,

Can you please explain in more detail what you mean in the following two sentences.  "Use at least 5 values for each continuous key parameter." and
"Use at least 5 replicates for each scenario."

[Alan Isaac]

BehaviorSpace

BehaviorSpace allows you to specify experiments where you vary the values of the variables that are set by sliders.  I want you to consider at least five values for each slider.  Once all the values are set, we have a scenario.  BehaviorSpace allows you to run each scenario multiple times.  Each run is called a "replicate" of that scenario.

Please ask if you have further questions,
Alan Isaac

[Alan Isaac]

UML

Just a reminder that we had this required reading during week 2:
https://subversion.american.edu/aisaac/notes/uml.xhtml

[Natalie Chambers]

When creating the UML decision node I noticed that the notes say that if the turtle is unhappy it is supposed to "move to "nearest" empty square that meets demands" however the code just moves the turtle to a random square without asking it to determine the conditions of that square. When creating the UML and going through the project should we follow what the notes say or what the code says?

[Alan Isaac]

Great attention to detail!  You may do either; just be clear about your choice.  (I have a very slight preference for you to follow the notes rather than the sample code.)

Alan Isaac

[Natalie Chambers]

Excellent thank you. Just wanted to post about a threshold I just discovered on the Segregation model. While playing around with the %-similar-wanted with a constant number of agents at 1970 (randomly selected on the slider) there seems to be a 75% threshold for which players can become segregated (at least in a reasonable amount of time). At 75% the players become segregated in around 147 ticks where as at 76% they continues to be moving well past 1000 ticks. This also applied when varying the numbers slider. I would be interested to discuss further why this is the case. I have a feeling there is a mathematical explanation that perhaps when they prefer more than 75% similar neighbors there is no possible combination of neighbors to satisfy these preferences but it would be interesting dive into this a bit. 

[Natalie Chambers]

Sorry to post yet again but I have a question about converting our data outputs from BehaviorSpace into a graphical representation (it is more of an excel question but it has really held me up in completing the project). When the data is exported and I open it with excel and attempt to create 3-D surface graphical representation I am unable to get the numbers to correctly align for proper graphing. It seems the numbers need to be arranged like:

  a  b  c  

x 1  2  3

y 3  2  1 

z 2  3  1

but when we get the output I am not sure how to get them arranged like this. I can transpose the x and y axes but how do you fill the z axis in the center. Would appreciate some help!

Thanks

[Alan Isaac]

3d Plots

Producing 3d charts in Excel is a pain unless you use an add-in.  It is also not (!) required for this assignment.

That said, here is a good way.  Select the x, y, and z columns in your spreadsheet.

Save them as a CSV file, say "mydata3d.csv".

Use (free and open source) gnuplot to plot them.

set datafile separator ","   #assuming a comma-separated values file
splot 'mydata3d.csv' using 1:2:3 with points palette pointsize 3 pointtype 7

You will probably want to save the file in PNG format for insertion into a document.

[Adaora Isichei]

Hey Natalie,

You may want to vary one variable and export your data as a table and not a spreadsheet.

[Adaora Isichei]

Hey Natalie,

I also observed the same and have to agree with you. Unfortunately, it seems as though netlogo's coding capacity limits the illustration of segregation at 75%.

[Alan Isaac]

I am not quite sure of the intent of this comment, but I want to be clear: NetLogo is a full-fledged programming language, and the results you see running NetLogo models you would also see if you implemented the same model in another language.
Alan Isaac