Repository files navigation

snahelper provides a set RStudio Addin for social network analysis.The main addin is theSNAhelper which provides a simple GUI to docommon network analytic tasks and visualize a network withggraph.

The second addin, calledNetbuilder allows you to quickly build smallnetworks with a small “canvas” to draw on. The network can be exportedas an igraph object at the end of the session by clicking on “Done”.

The third addinNetreader is meant to facilitated the import of rawnetwork data. It provides a GUI to easily read network and attributedata and combine them to an igraph object. The underlying code of theimport procedure is shown at the end. This should help users to learnimporting data themselves.

The fourth addinComponentlayouter allows to layout networks withseveral components manually by placing them on an empty canvas.Components can also be rotated and resized. After finishing the session,the layout is saved as x and y vertex attributes.

Code to reproduce the used network.

library(tidyverse)library(igraph)url<-"https://raw.githubusercontent.com/evelinag/StarWars-social-network/master/networks/starwars-full-interactions-allCharacters.json"df<-jsonlite::fromJSON(url)df$nodes$id<-1:nrow(df$nodes)df$nodes$display<-df$nodes$namedf$nodes$name<-df$nodes$id-1g<- graph_from_data_frame(df$links,directed=F,vertices=df$nodes)V(g)$name<- V(g)$displayg<- remove.vertex.attribute(g,"id")g<- remove.vertex.attribute(g,"display")g<- delete.vertices(g,which(degree(g)==0))V(g)$display<- ifelse(V(g)$value>75,V(g)$name,"")V(g)$colour<- ifelse(V(g)$display=="",NA,V(g)$display)g

Netreader should be pretty selfexplanatory. The first two tabs allowyou to import raw data (edges and attributes). Make sure to specify filedelimiters, etc. according to the shown preview.

Using theNetreader should comes with a learning effect (hopefully).The last tab shows the R code to produce the network with the chosendatawithout using the Addin.

The network will be saved in your global environment once you click“Done”.

Highlight an igraph object in your script and selcet theComponentlayouter from the RStudio addin menu. Components are added byclicking on the canvas and can be resized/rotated and re-positionedusing the buttons at the bottom. The final layout is stored as x and ycoordinates in the igraph object.

# developer version#install.packages(remotes)remotes::install_github("schochastics/snahelper")#CRAN versioninstall.packages("snahelper")

To work properly, you also needgraphlayouts, which addsnew layout algorithms.

install.packages("graphlayouts")

To use the main addin, simply highlight a network in your script andselectSNAhelper from the Addin dropdown menu.

The layout tab allows you to choose from all implemented algorithms inigraph and some layouts fromgraphlayouts. The default is a stressbased layout and also the recommended choice. Seemyblog foran explanation. In the tweak section you can move individual nodesaround. Choose the node from the dropdown menu and click on its newposition in the plot.

The Node Attribute Manager shows all existing node attributes in asortable table. In addition, you can calculate some new ones (centralityand clustering). The functions automatically choose the right version ofindices, depending if the network isdirected/weighted/undirected/unweighted.

This is where you can style your nodes. You can either do it manually,by choosing a color/size for all nodes together, or based on anattribute.

Same as Node Attribute Manager but for edges. So far only shows existingedge attributes.

You can style your edges here.snahelper automatically detects if yournetwork is directed and adds arrows if the network is directed. Theother options are similar to the nodes tab. The snahelper automaticallychooses the appropriate edge geom. If multiple edges are present, itusesgeom_edge_parallel0(). Otherwisegeom_edge_link0().

The result tab shows the network in its full size. If you are satisfiedwith the results, hit the Done button and the R code to produce the plotis automatically inserted in your script. Or you can directly save theresult as a png file.