As two core roles in film industry, composers and producers are independent but also rely on each other. Do they have some network? Who is the most popular composer from producers’ perspective? Which group of people is the center of these people?
This network contains the collaboration of 40 composers of film scores who created the music in the period 1964- 1976 and the 62 producers who produced a minimum of five movies in Hollywood, 1964-1976.
I did not get some specific inspiration about this work. Just the general imagination from class of the a group of people, and find the core people and the center social circle. Also the main cluster of them and the people who are in the edge of the network.
CASOS: hTo be honest, I did not have this clear topic when I looked for the data at first. I was more like looking around all data from the three websites on slides and trying to find some interesting ones(also easy to understand.) After five or six times of failure because of the inappropriate size of data or nothing to found in the results, I got this film producer and composer network data. That was pretty quickly compared to my previous work. I got this data from CASOS.
OpenRefine: useful and friendly tool to manage my data like always.
Gephi: main visualization tool in this work, to help me run the different statistic and build the final network. Easy to crush, and cannot go back, so saving every time before changing.
To make the data could be recognized by OpenRefine and Gephi, I imported the data without the first hundreds of lines, like some explanation. However, in this way, the network graph I got from Gephi was only with their numbers instead of their names. The only way I came up with was to transfer those by coping and pasting myself. I did not save the screenshot the the “before” version, and here is the “with name” version.
I run the three statics to make it more like a cluster and make sense to me. Then, I changed the color of the main group of people to make it more clear. Also, I filtered the graph by the 5 and 8 to explore the network of their core people.
Results & Reflection
Here is the final visualization I got. I chose the people who have more than five network to colored by different color of their cluster. Also their name circle would be sized by their influence. As you can see, Goldsmith, J, Schifrin, L, Bernstein, E are the three most “popular” people in this visualization. They are all composers. The reason why the most popular three people are composers instead of producers, my guessing would be that the composers are easier to make songs for many different film. Making songs would use much less time than run and produce a film. Therefore, it make more sense for me that the popular composers would have lots of chances to write songs, then have network on this network.
Also, if you observe it carefully, you will find that the people connect with composers would all be producers due to they could only have connection with the other type of artist.
The two graphs below are with the filter that only keep people who have more than five(left one) and eight(right one) connections with others. With left graph, we might feel like the Goldsmith, J is the biggest cluster, but the right one let us feel the other seven artists are the core cluster which have a more close and wide connection.
This colorful one below is the graph colored by their modularity class after the “more than five connections filter.”
There are one thing I really want to improve, which is the classification of composers and producers. I feel confused to distinguish these two types of artists. However, at the same time, I want to add different colors to differentiate their clusters. Or I could use different Gradient colors? I don’t know, but that was definitely the point I want to change if I could do it again.
Also, Gephi is even harder than Carto to learn and use. I carefully listen the knowledge on class, but I still felt difficult to use it fluently. At the same time, it does make sense to me that every tool is harder than the last one lol. I think I need to practice more to use Gephi. Making network visualization is more like an exploration compared the first three lab work. We never know what will come up and if it is meaningful. Be patient on it is the most important thing that I learned this time.