Visualizing Jazz Musicians Network using Gephi


Visualization

Introduction

During the class, I found it very interesting to know that we can visualize our social network connections using Gephi. Since Facebook is so widely used by everyone, I wanted to try creating one visualization for myself. While searching, I got to this article http://allthingsgraphed.com/2014/08/28/facebook-friends-network/ where I found about Netvizz application. Upon using the app I found out that it doesn’t provide the dataset for your connections anymore, it only provides datasets for your search, page likes and shared data. So, I was just playing with some datasets from Gephi-wiki and I came across Jazz Musicians Network and decided to go with it.

Methodology

For inspiration I looked at the famous Linked Jazz network visualizations. It made me curious as to how so many artists are linked together, what could be the things that are connecting them together. I wanted to try and replicate the same for my project.

linked_jazz_full

lj-network-jane-jarvis

First, I downloaded the data from Gephi-wiki which was in the .net format. This format was accepted by Gephi so I didn’t have to change anything. Next, I had to change the layout of the data and tried few of the layout options to get the same linked jazz viz but it didn’t work out. So, I searched for some social network visualizations and found these.

facebook_nonames1gephiheader

Initially I started with Force Atlas 2, where I ran the statistics for Average Degree, Modularity and Graph density. With this data I was able to adjust the size of the nodes and also changed the colors of the visualization according to the modularity class. I was still not satisfied with the layout, I played around with the properties of the layout (gravity, scaling) but it did not work. So, I went with Fruchterman Reingold which actually made sense, all the clusters and connections were now properly visible. One could assume seeing the visualization, how these musicians are separated by different locations (east coast & west coast) or are connected by the same jazz genres they play.

untitled

I also wanted to add the names of the musicians to the visualization but the dataset I used did not have their names. I tried making the hypergraph of this viz but was not able to do it.

Future

If I could find the same dataset with names, I hope to create the same Linked Jazz Network visualization with name and pictures.