What We Studied

We decided to focus on the research questions: What attributes and background lead to the most connections on Instagram between our 5 groups of friends?

We completed this by looking into a variety of attributes including:Year, Home State, Major, Sex, Club Involvement, Greeklife, and Citizenship

Why

We chose this topic because of our interest in the egocentric networks each of us has at GW. Knowing that students connect in a variety of different ways, we wanted to better understand the sources of acquaintances, based on our 5 egocentric networks.

How

Each group member identified 12 of their closest friends. We then looked at the connections between the 68 people identified total. With this information we performed some exploratory and in depth analysis, here is what we found:

Plot Of Network

Network Properties

## The size of the network is 797
## The density of the network is 0.1650104 which means that network is not dense
## There is 1 component in the model, and everyone is part of that component. This shows that everyone in network could reach any other person.
## The diameter of the network is 4 and that is the longest path between two individuals. The diameter helps us understand the length of our relations and that each person is connected to other through a path of 4 or less.
## The average path length of network is 2.214493 which shows average shortest path between two nodes.
## 0.9335006 percent of students are following eachother mutually on Instagram

## Based on the histogram, the network seems to be like a real-network because it is sparse.

Plots with Shapes and Colors

## For the network, Business students are tomato, Science students are blue, and IA students are black, and others are white
## For the plot of the network, square is if person is female, and circle is if person is male
## This graph plot shows that there is a concentrated businness group and then dispersed around that are other majors
## After we plot the graph based on the sex and major of students, we could see that Female is the dominant gender and Business is the dominant major. The technique we used for this problem is adding the attributes and dividing them into different colors and shapes. It is reasonable that females are the dominating sex because GW has a higher female student ratio than male students. Business is the dominating major is also understandable because DNSC 4233 is a class under the business school, which makes sense that the tomato color is widely shown in the graph.

Clusters and Communities

## This shows that there are four communities within the network
## The technique we used for this section is clustering. After using the cluster_walkstrap function, we found out that we have four different communities within the graph, highlighted in red, green, blue, and purple. Each community seems to have densely connected ties within the cluster. The red edges indicate a connection between a member within the group and a member outside the same cluster. However, the modularity score is 0.37, which is less than 0.5. A low modularity score demonstrates more edges connecting members from two different communities, and the communities have a lower individual density. The low modularity score makes sense because all the students are from GW. Students could communicate outside their communities because they could be involved in different GW communities and meet other people.

Assortivity

Triad Census

##  [1] 31188  2417 16922    22    20    27   396   392     1     1  2221    17
## [13]    16     4   181   915

## Based on this graph, we can see that the most popular triad is type 1 which is no connection between three people. The second most popular is type 3 which is a mutual following between 2 nodes. The third most popular is type 2 which is a one way follow. Lastly, the fourth most popular is that a and b follow each other, b and c follow each other.

BoxPlots

## Now we will take a look at boxplot distributions of degrees. These will help show which subset of students are better connected.

## This shows that based on degree, the business school students are more central

## This shows that based on degree, the seniors and sophmores are almost equally  as central

## This shows that based on degree, students who are not international students are more central

## This shows that based on degree, students involved in Greek Life are on average more central

## This shows that based on degree, females and males have same average degree, but the distribution for females is larger

Conclusion

In conclusion, there are very interesting information that can be gleaned from the network, that has been illustrated above. But overall, A GWSB 4-th year female who is part of Greek Life is deemed to have the most connections out of our 5 ego-centric networks.