Patterns in Oscar Movies
Authors and Affiliations:
The data visualized here was taken from the InfoVis 2007 datavisualization contest (conference info). However, we did not submit this data in the actual contest.
To render the XML-formatted data set, we used Processing, a Java-based prototyping tool created by Ben Fry and Casey Reas.
With the large amount of data we had to work with, our first task involved minimizing data se and finding a spefic segment of data we would effectively use.
We thought it would be interesting to examine the relationships between actors who have won Oscars, the directors they have worked with and all the other actors they have worked with. Using Processing and the ProXML Library by Christian Riekoff to load the information, the XML data is looped through and parsed. The Oscar winning actors were designated to the middle ring, the directors they had worked with to the center ring and the other actors they have worked with to the outer most ring. Bezir curves were used to draw the relationship between each individual between each circle. These curves allow us to examine the relationships between Oscar winning actors and their peers.
- Image 1.1:
With over 10,700 rated movies found in the data set, we decided to narrow in on Oscar-winning movies alone to demonstrate the relationship between Oscar-winning Directors to Oscar-winning actors. Then show the relationship and frequency between Oscar-winning actors and non-Oscar-winning actors in Oscar-winning movies. To do this, three circular paths were created to differentiate directors (center-most circle labels) from Oscar-winning actors (center circle labels) and non-Oscar-winning actors (outer-most circle labels). Curved lines were programmed to draw the relationships between directors to Oscar-winning actors and between Oscar-winning actors and non-Oscar-winning actors. The curves help to better depict the individuals of the relationships. Also, each line is semi-transparent, so that when multiple relationships between individuals occur, the color intensity of the line increases.
The result is a wheel-like diagram with seemingly chaotic spokes within it. While the names of directors and Oscar-winning actors can be difficult to read, one can clearly see specific patterns in the relationships between individuals. One of the interesting discoveries we found here were the few non-Oscar-winning actors who have worked with multiple Oscar-winning actors multiple times.
- Caption for exhibit:
Using Oscar-winning movies only, a diagram is programmed to draw the relationships between Oscar-winning directors (inner-most circle) and Oscar-winning actors (middle circle) and between Oscar-winning actors and non-Oscar-winning actors (outer circle). The yellow curves represent two people who have worked together on a movie.
Our idea for the second task builds directly upon our first task. We decided to investigate the relationships between non-Oscar-winning actors to investiage relationships with a much larger amount of people. At first we attempted drawing direct links from actor to actor based upon whether or which actors have been in a movie together. The initial results revealed and illegible, almost solid white mass. To accomplish the task of bringing the lines outside of the circle, we used Catmull-Rom spline curves.
- Image 2.1:
Taking the diagram constructed in Task 1, we decided to try to decipher relationships between non-Oscar-winning actors. That is, for each non-Oscar-winning actor, what other actors have they worked with. The initial results revealed a mass of chaos of white lines that was impossible to read. Retaining the lines, we decided to bring all lines around the outer rim of the circle except for a randomly selected few actors. From these few actors, one can see the strong relationships they have with a certain set of other actors. One could assume this demonstrates networking capabilities of the actor or perhaps the close net of professionals they work with repeatedly.
Acting is no different than any other profession in the sense that, certain people work well with others. Often times a particular director or actor will work multiple times with a certain actor. Sometimes as many as five or six, or even more, actors will work together on one movie then together on another. This is demonstrated by the intensity of certain links between actors; pairs of actors whose link is more brilliant have worked multiple times with each other.
- Caption for exhibit:
The white lines are relationships between all non-Oscar-winning actors. Allowing all the lines to cross over the circle resulted in illegible chaos without significant patterns. However, by pushing all lines except for a few randomly selected actors, one can get a glimpse of the relationships actors have to one another, often working repeatedly with the same group of actors.
For the third task we decided to investigate the relationships between genres. We considered the same set again for this task; just actors who have won Oscars for their performances. If a movie one an Oscar of any type, the size of genres associated to that film was incremented. To further demonstrate the distribution of Oscars to genres, we included all the genres, even those who receieved no Oscars, however these circles are left unfilled.
- Image 3.1:
In this diagram, we wanted to exemplify the type of genres in the data set and demonstrate the Oscar-winning categories that prevailed against other genres. For each Oscar won, the genre circle would increase in size. Circles without a fill represent genres with no Oscars won. The drama genre clearly has the most actor Oscars to its name with the crime, thriller and romance genres vieing for second. It is interesting to note that the actor Oscars seem to only be given out to about half the genres. This makes sense when one considers what genres did not have any actors receive an Oscar.
- Caption for exhibit:
For each Oscar won, the genre circle would increase in size. Circles without a fill represent genres with no Oscars won.