Here you can find various things that might help you: R code (usually related to networks), presentations, and other small things that I think could help someone.
R code (usually related to networks)
Convert network data from a node list to an edge list. A node list stores all relationships for a given node in one row of data with N columns. It can be an efficient way to store or structure data and you will often encounter network data formatted this way (e.g., from online surveys, exports from citation databases, etc.) Most R network packages cannot easily read data from a node list. The function converts the node list so that the N relationships (pairs of nodes) are stores in N rows and two columns, which is easy to import into most R network packages. (This is a small R function and the file is saved as a .txt as wordpress will not host .R files)
Convert network data from an edge list to a node list. An edge list stores all relationships for a given node in N rows of two columns (where column 1 = node 1 and column 2 = each of its relations). You are probably less likely to want to go from an edge list to a node list, but you might want to. The function converts the edge list so each node’s relations are stored in a single row. It allows you to extract two columns of data from a larger data frame and to customize the output column names. (This is a small R function and the file is saved as a .txt as wordpress will not host .R files)
Calculate two-mode network centralization. Do you stay up late at night worrying about how you will calculate two-mode network centralization? If so, this little R function will help you sleep. A two-mode network (also known as a bipartite network) has two sets of nodes (e.g., people and events) and relationships can only exist between sets. Centralization is a graph-level measures describing the distribution of network relationships. Most network packages calculate centralization for one-mode networks, but not for two-mode. This function operationalizes Everett’s and Borgatti’s (2005) equation 4.10 for two-mode centralization. I used it in Sayles and Baggio (2017 PNAS). Maybe you will find a use for it too? (This is a small R function and the file is saved as a .txt as wordpress will not host .R files)
Citation: Everett M, Borgatti SP (2005) Extending centrality. Models and Methods in Social Network Analysis, eds Carrington P, Scott J, Wasserman S (Cambridge Univ Press, Cambridge, UK), pp 57–76.
Calculate multiplex participation coefficient and z-scores. I know what you are thinking … your birthday came early this year: you can now calculate multiplex participation coefficients (Pi) and z-scores with this function. A multiplex network allows for multiple kinds of relationships between any pair of nodes. Pi measures how evenly a node’s edges are distributed among different categories of relationships. Z-scores measure a node’s total number of relationships among all categories (and normalizes them to account for network size). This equation operationalizes Battistonet al’s (2014) work as an R function. I used it in Debortoli et al (2018 ERL). (This is a small R function and the file is saved as a .txt as wordpress will not host .R files)
Presentations
Different ways that network science can be used to aid environmental research and management. This is a short (6 min) presentation that I gave titled: Networks are more than just stakeholder maps: everything you didn’t even know you wanted to know about network science in six minutes. It is from a lightning talk series at the US EPA Social-ecological science exchange, 24 April 2019.
Other items: