|
From: | Andrew Papachristos |
Subject: | [igraph] two-mode to one-mode with event types/attributes - thank you (Andrew Papachristos) |
Date: | Tue, 15 May 2012 12:36:36 -0400 |
Moses and Gabor, Thanks for the quick help....I haven't tried the last part using 0.6 yet, but the rest seems to work smoothly! Thanks. Moses, to answer your question, I'm looking create a series of networks among offenders/criminals for an entire city using multiple events: arrests, observations, visits, citations, and more. In classic two-mode --> one-mode fashion, I'm assuming there is a tie when two people have the same event code (and there are exact event codes for a each arrest, observation, etc). Thus two people arrested in same incident/event share an event code and, by extension, a "tie". There are also time stamps on these data covering nearly a decade. Because I'll most likely be aggregating to a month/quarter, this does indeed mean two individuals can have multiple events occurring in same time period--i.e., they were both observed together and arrested together in the same month. So, in this case the event indexation = the type of event (arrest, observation, citation, etc), and the event time is the actual date it said observation occurred. I'm still in the early stages of this project, but I'm headed in a direction that will want to compare the different event/network types (arrest v. observations, say) at different time points (before or after different interventions, for instance). best, andrew Andrew V. Papachristos Robert Wood Johnson Health & Society Scholar ~ Harvard University Assistant Professor (on leave) ~ UMass, Amherst - Dept. of Sociology (413) 545 - 0443 |
[Prev in Thread] | Current Thread | [Next in Thread] |