README file for Flow Through Centrality

The file Code.zip is a compressed file containing all of the data and R markdown files of the analyses done for JDS220-011, "Flowthrough Centrality: A Stable Node Centrality Measure".

Once unzipped, the folder "Code" contains three sub-folders "Analyses", "HijackerData", and "Simulations".

# Analyses
The "Analyses" sub-folder contains final R Markdown files for the figures and data that were used in the manuscript. Each markdown file contains R code for importing the data into R, applying the Nashimoto-Wright test to each simulation scenario, and obtaining the Type I error. There is also R code for creating the line graphs shown in the manuscript.

# HijackerData
This folder contains the R markdown files and .csv files for the analysis of the network of 9-11 hijackers using the six centrality measures under comparison. The real network was perturbed by removing 3%, 6%, 9%, and 12% of the links in four different experiments (one experiment for each link). For each percentage perturbation, the various centrality measures were calculated 10 times. The purpose of this experiment is to show the stability of flowthrough centrality and flow-between centrality in comparison to the other measures. The original Excel files are also contained in a subfolder. They are the same as the .csv files except for the file format.

# Simulations
The Simulations sub-folder contains 20 further sub-folders, each one for a different simulation scenario. The parameters for the simulations were the number of nodes (50, 100 or 150), the number of degrees (6 or 10) and the number of perturbations (3, 6, 9, or 12). The number of perturbations is the number of links between nodes that were removed. Folders are named NodesADegBPertC, where A = the number of nodes, B= the number of degrees, and C = the number of perturbations. For example, the folder "Nodes50Deg6Pert3" contains the analysis files and data pertaining to the scenario with 50 nodes, 6 degrees, and 3 perturbations.

Within each sub-sub-folder of the Simulation directory are 8 more files. Five of the files are .csv files, and they contain the data on the measures of centrality for each simulation. There is a file for betweenness, for closeness, for flowthrough for flow-between, and for stable betweenness. 

Within each directory is also a .txt file, which contains R code for entering the .csv files into R and wrangling the data files in order to accomplish the analysis set forth in the R Markdown document. The .html document is the output when the R Markdown document is compiled.



 


