NodeXL and My Little Secret

I have to share a secret with you: I am really not too great with technology.

Making this blog was difficult for me at first, and I needed a lot of help from one of my friends, who blogged about her study abroad experience in Spain last spring (thanks, Mia). I still can't do all the fancy stuff or get all the details of formatting exactly as I want them, but I have the basics down now, at least.

And then I was told I would have to do a project using NodeXL, a program attached to Excel that allows you to import data from different social media and look at connections between them. I was afraid, very afraid. My professor was sitting in front of me, going on and on about all the exciting things that I could do and see using NodeXL, and all I could think to myself was: "It's not gonna happen like you want it to, that's for sure." I'm horrible with things like this.

Make it interesting, colorful, exciting, insightful--he tells me. Here's what I came up with.

A grey blob.

I was thrilled. I was not expecting it to turn out so well.

I come back the next week to meet with my professor. He is hesitant about what to say for this grey blob that I am so proud of myself for making. Maybe I could change the colors of the nodes? (How do you do that? What are nodes?) How about I change the sizes and shapes and explore the program a little bit more to see what I can come up with? (Can you please tell me exactly what you want me to come up with? I can't be creative when I am having so many technical difficulties--on the fault of my own, not the program.)

Over the next two weeks, he gives me a crash course in the basics of how to adjust certain things about the program, and a few articles to read to give me some new ideas. Suddenly my nodes are turned into images, then transformed into colorful dots of all shapes and sizes. The edges go from opaque and clumped to transparent and spread out so they are easier to see. I am actually starting to make some observations. Here are some of the "round 2" graphs. By the way, you can click on any of these to see them fullscreen.

My professor tells me he likes the third one the best and gives me some pointers about how to edit it further. Let me stop and talk a bit about what the project is about, before I reveal the final product.

The black node is me. Every other node represents one of my followers on Twitter. I started by importing all of them, which was a challenge because I overloaded the system a few times and it basically told me to "come back later." I was a little over-ambitious. By the time it was done, I had myself plus 102 other people in a neat little row on an Excel sheet. I could click back and forth between "vertices" and "edges" (something I didn't figure out until week 2). There's a very long string of information for each person or each connection between people, but I only manipulated the first few columns.

I changed the color, shape, and size to represent my different groups. I found that I had a large, tight-knit group of people living in Athens and also involved with a program called Young Life. I currently live in Athens for most of the year and I am involved in Young Life, by the way. On the opposite end of the spectrum were generally high school friends not involved in Young Life. There were some people living in Athens who were not involved with Young Life, and some people involved in Young Life who did not live in Athens, but these were the exceptions to the rule. Here is the order of colors representing people most involved in Young Life, to people least involved: blue, green, yellow, orange, purple, red.

As you can see, color matters, but size matters more. Large nodes are Athens people and small nodes are not. Squares represent Twitter accounts who are not real people, but circles represent real people. In addition, I made the edges more transparent so they were easier to see. For more information about my findings from this project, click here. I won't go into those now.

The crazy thing was, once I really got going, it was actually a lot of fun. Maybe I was just being (literally) egocentric and enjoyed learning about my own network, but the program became a lot less intimidating than it was when I started. I began to figure out how to manipulate certain things after a while, and I actually enjoyed it. Here is the final graph that I created. A lot better than the grey blob.

P.S.   You'll keep my secret, won't you?

Tweet, Twitter, Tweetest

Check it out! I made a Node XL graph of my Twitter followers! This is the final product, but I went through a lot of "drafts," modifying it a little each time. That's me in the middle. Go ahead and click on it to see it better.

One thing that I noted about my followers was that they fell into two main groups: people I know from Athens, and people I met prior to moving to Athens. Additionally, they are either involved with Young Life in some way, or they are not associated with Young Life.

Young Life is a Christian outreach program that places "leaders" at a specific high school. Different programs are run for students of each high school, and in this case, I am calling the high school students who attend these events "participants."

You probably wouldn't be able to make sense of this unless you knew what the colors, shapes, sizes and labels meant. The colors and the labels mean the same thing for each node. There are 6 categories of colors/labelled nodes: Young Life leader (blue; "l"), former Young Life leader (green; "fl"), Young Life participant (yellow; "p"), former Young Life participant (orange; "fp"), associated with Young Life in another way (purple; "as"), and not associated with Young Life (red; "na").

There are two different sizes for nodes, not including my own. The larger nodes represent people I met in Athens, and the smaller nodes represent people I did not. Finally, most of the nodes are circles, but some nodes are squares because they do not represent actual people. An example is "@YLencouragement," which is not a real person and is therefore represented as a square.

As you can see from the chart, most of the people I met in Athens were people involved with Young Life in some way. There are 5 exceptions: 3 of these are my roommates, 1 is my roommate's boyfriend, and the 5th is an Athens High School (AHS) student who was on the cross country team that I coached last fall. This AHS student is near a lot of large, yellow nodes because these nodes are participants in Young Life who also go to AHS.

All green and purple dots are near the large "clump" of Young Life involved, Athens-located nodes. However, there are 5 small blue nodes that are farther away. 4 of these are people who attended my own high school, and now are leaders at locations other than Athens. The 5th node was actually my own Young Life leader when I was in high school.

I can't forget to mention a very unique node on the graph is the only node connected only to me. This is my cousin from Dayton, OH. (I am from a town between Dayton and Cincinnati.) Everyone else is at least connected to one other person in the graph besides me. I, of course, am connected to everyone (I'm so popular!)

The Athens-located/Young-Life-involved clump generally looks as expected, with many current leaders in the middle, surrounded by former leaders and others associated with Young Life. Most of the "others associated" are people who went through the training program to become a leader, but did not actually become one. The small purple node represents a woman married to a Young Life leader from my high school, and she helps out with events although she is not a leader.

There really aren't any small nodes that are connected to larger ones (who are real people), with two exceptions. One of the people I went to high school with (small, blue node) is connected to one of the leaders from Athens (large, blue node), and I didn't even know they knew each other until I made this graph! The only other connection is between one of my non-Young-Life college friends (large, red node) and my boyfriend (small, orange node); they met when he came to visit me here in Athens.

Seems like what's most important here is the size of the node, second only to the color of the node. Node sizes almost do not mix at all, but similar colors mix fairly often. The shape of the node hardly mattered at all; square nodes of the same shape and color lay together, but square nodes of different shapes and/or colors did not.

If you want to read about my experience and some of my failed attempts using NodeXL, click here.