—
title: “Example 2”
author: “sheng huo”
date: “April 4, 2018”
output: html_document
—
“`{r}
library(knitr)
setwd(“/Users/NOAH/Desktop/CSC 495/week2/ex2/ex2”)
read_chunk(“example02.R”)
knitr::opts_chunk$set(echo = TRUE)
“`
## Working with different types of data
### Step 1: Load the necessary libraries and data
“`{r C1, results=”hide”, warning=FALSE, message=FALSE}
“`
Loading the Southern Women network (Davis data). GraphML format.
“`{r C2}
“`
### Step 2: Examine the bipartite network
Verify that it is bipartite.
“`{r C3}
“`
Plot using bipartite layout (better if rotated)
“`{r C4}
“`
Look at labels and types
“`{r C5}
“`
So, FALSE is people
### Step 3: Create projection
Create person-person projection
“`{r C6}
“`
Plot the projected graph
“`{r C7}
“`
Plot the distribution of edge weights
“`{r C8}
“`
### Step 4: Filtered version of the network
Create a new network removing the edges of weight 1 and 2.
“`{r C9}
“`
Plot the weighted degree distribution of the new network
“`{r C10}
“`
### Step 5: Ego networks
Extract the ego networks for Laura and Sylvia
“`{r C11}
“`
Plot side-by-side use layout_as_star
“`{r C12}
“`
Compare the two ego networks by weighted degree. First create data frame with the right structure.
wdeg, ego
10, Laura
24, Sylvia
34, Laura
“`{r C13}
“`
Then plotting with ggplot is straightforward.
“`{r C14}
“`
### Step 6: New example: loading edge and attribute data
Load CSV files for edges and attributes. stringAsFactors is
false because we can’t use factors as node attributes.
“`{r C15}
“`
Convert to a graph from the two data frames. Note that the
names of the vertices in the edge data frame have to match the
first column of the vertices data frame.
“`{r C16}
“`
Plot
“`{r C17}
“`