Along with recent shifts in the Sociology of Culture towards relational techniques is the use of the correlation network. Instead of examining the answers to survey responses themselves, these approaches look at relationships between questions and try to take meaning from structural properties of the whole. I also used some of these techniques for the GHTC 2016 conference [1] with Lee exploring USAID data from Guatemala. The results appear in our paper [1], but the true inspiration comes from the work on statistical methods for gene co-expression [2]. One particularly exciting work in sociology [3] also tries to explain the structure of political beliefs using these networks.
I plan to do further work using these techniques, so I created a python library for anyone interested.
WCNA Python Library GitHub
Continue reading “Weighted Correlation Network Analysis Python Library”