Tutorial

SimpactPurple is a work in progress. This tutorial will walk a user through running a SimpactPurple simulation, changing some features, and generating some output.

Run a Simulation

Create a community with default settings and run the simulation:

>>> import simpactpurple
>>> s = simpactpurple.Community()
>>> s.run()

We can visualize the result with the GraphsAndData module:

>>> import simpactpurple.GraphsAndData as GraphsAndData
>>> GraphsAndData.age_mixing_graph(s)
>>> GraphsAndData.sexual_network_graph(s)

Or investigate agents and relationships ourselves:

>>> some_agents = s.agents.values()[0:10]  # agents are stored in a dictionary, so need to use values method
>>> some_agents
[<Agent.Agent instance at 0x00000000081377C8>, <Agent.Agent instance at 0x0000000008137808>, <Agent.Agent instance at 0x0000000008137848>, <Agent.Agent instance at 0x0000000008137888>, <Agent.Agent instance at 0x00000000081378C8>, <Agent.Agent instance at 0x0000000008137908>, <Agent.Agent instance at 0x0000000008137948>, <Agent.Agent instance at 0x0000000008137988>, <Agent.Agent instance at 0x00000000081379C8>, <Agent.Agent instance at 0x0000000008137A08>]
>>> some_agents[0].attributes  # access attributes of a particular agent
{'TIME_ADDED': 0, 'NAME': 0, 'TIME_REMOVED': inf}
>>> some_relationships = s.relationships[0:10]  # all relationships stored in a list
>>> some_relationships
[(<Agent.Agent instance at 0x0000000008137B88>, <Agent.Agent instance at 0x00000000081377C8>, 0, 16), (<Agent.Agent instance at 0x0000000008137F08>, <Agent.Agent instance at 0x00000000081377C8>, 0, 18), (<Agent.Agent instance at 0x0000000008138608>, <Agent.Agent instance at 0x0000000008137F48>, 0, 17), (<Agent.Agent instance at 0x0000000008137908>, <Agent.Agent instance at 0x0000000008137E88>, 0, 17), (<Agent.Agent instance at 0x0000000008137948>, <Agent.Agent instance at 0x00000000081386C8>, 0, 10), (<Agent.Agent instance at 0x0000000008137BC8>, <Agent.Agent instance at 0x0000000008137848>, 0, 18), (<Agent.Agent instance at 0x0000000008137988>, <Agent.Agent instance at 0x0000000008138048>, 0, 17), (<Agent.Agent instance at 0x0000000008137A08>, <Agent.Agent instance at 0x0000000008137AC8>, 1, 14), (<Agent.Agent instance at 0x0000000008138D08>, <Agent.Agent instance at 0x0000000008137B48>, 1, 15), (<Agent.Agent instance at 0x0000000008137EC8>, <Agent.Agent instance at 0x0000000008138808>, 1, 18)]
>>> some_relationships[0]  # a relationship is stored as a tuple (agent1, agent2, start, end)
(<Agent.Agent instance at 0x0000000008137B88>, <Agent.Agent instance at 0x00000000081377C8>, 0, 16)

We can change some of the default values, including initial population and number of years to simulate:

>>> s = simpactpurple.Community()
>>> s.INITIAL_POPULATION = 100
>>> s.NUMBER_OF_YEARS = 30
>>> s.run()

You can also customize things like agents preferred degreed (desired number of partners – DNP) or the gender ratio:

>>> s = simpactpurple.Community()
>>> s.DNP = lambda: 2  # every agent has a DNP of 2
>>> s.SEX = lambda: random.random < 0.25  # prob[agent == male] = 0.25
>>> s.run()

For more examples check out the example scripts in the /bin folder in the home directory.

More Advanced Use

Currently changing the hazard of forming a relationship requires changing the source code in __init__.py. Future releases will make this easier, but is not possible at this point.