Palladio is a web based application that allows the user to create a series of cluster graphs that are based on different networks described in a series of data. These networks represent different types of relationships such as those between different people, different places, different time periods, or different objects that are in some ways connected to each other. Through using the application, the user is able to reveal the relationship of these networks and reveal their common traits. When using Palladio to create and analyze different networks contained in a data set, the user is asked to select two different categories that were used in that data set. One category is selected as the source and the other is selected at the target.
For my personal work with Palladio, the data I used came from the WPA Salve Narratives, which was a collection of interviews of former slaves conducted between the years 1936-1938 as part of the Federal Writer’s Project of the Works Progress Administration. I did not look at the whole interview collection for my work, instead I focused specifically on the interviews conducted in the state of Alabama. I used Palladio to construct a series of different graphs that analyzed the different networks contained in the data set. There were two main networks that I looked at in conducting my analysis.
The first network was the relationship between the interviewer and variety of different targets such as the subject of the interview, whether the person they interviewed was male or female, what type of slave the person they interviewed was, and what topics were discussed in the interviews. I was looking to see if who the interviewer was affected any of these topics. I found that it just depended on which interviewer it was. For some interviewers there was a different relationship and for others there was no difference in the relationship.
The second network that I looked at was how a variety of different factors could influence the topics of the interview. The factors I looked at were what type of slave the person interviewed was, the age of the person interviewed, and where the person interviewed was enslaved. I found a similar result with this network as I did with the first one. Some of these factors did influence what the topics discussed in the interviews were. However, there were certain topics that did overlap across the different categories.
Through using Palladio I was able to discover a variety of different networks that are contained within the WPA Slave Narratives for the state of Alabama. These networks revealed some common trends between the different categories contained in the data set, that were revealed by creating cluster graphs in Palladio.
