Helen Keithahn and Emma Wasend
Progress
We have been observing and gathering pieces of data, such as old Carletonians, from Carleton’s online digital archives and historical resources found on the Carleton website. We have also visited an archivist in the Library’s archives and received some data on increases and decreases of enrollment rates throughout Carleton’s history. Recently, we were referred a contact from admissions on gathering more data about enrollment rates in past years.
Problems (and proposed solutions)
Our project has undergone a few changes since we first conceived of it. In the first stages of its being, we wanted to look at the atmosphere of prejudice on Carleton’s campus around the time of the Vietnam war, given the college’s unique relationship to Japanese students at the time. However, this proved to be too broad a subject and hazy a relationship to be described by the data that we had access to. After talking to Tom Lamb about our project, he supplied us with data and advice. While creating a project that would detail the reactions of the college at the time would be incredibly interesting, it was also unlikely that we would achieve this in so short a time span. Instead, he said, the clean data of enrollment statistics of college students at the time would be easier to identify some kind of relationship. We thanked him, and decided to broaden our question in span of time, and narrow it in specificity. What did enrollment statistics say about the war? How would we be able to identify what they were saying? In order to get a clearer picture, we decided to compare the statistics to similar numbers about the time of the first world war. It helped that Tom supplied access to these data as well. He recommended that we reach out to admissions to see if they had any data about the times as well.
A few of the problems that have presented themselves are the lack of context surrounding these data. We want to tell a story, not just show a difference between the enrollment during the time of the vietnam and world war 2. In response to this problem, we have been searching through the published editions of the Carletonian during that time, looking for references to the war, and specifically what was going on with Carleton at the time. Another problem facing us is the fact that the data is not the number of students who actually went to Carleton but the number of students who enrolled. This means that any story we tell or comparison we make will be based on the students who enroll, and any who did not actually attend will not be included. A simple solution, is to compare multiple years, as that gives a more comprehensive understanding of the number of students actually at the school.
Tools and Techniques
We are using excel, Voyant tools and (hopefully) Timeline JS
Deliverables
We will meet again with the Archivists tomorrow
Begin cleaning data after meeting
Have comparative data in presentational form by next wednesday
Have contextual data in presentational form by next wednesday
Emma and Helen,
It sounds like you have refined your project scope and gotten some good advice from the college archivists. As we discussed in class, I think a good solution to your problem with context would be to use sentiment analysis on a selection of Carletonian articles about each war (or, possibly, on certain relevant runs of the paper during the key points of each war). This would let you compare how Carleton attitudes toward the conflicts correlate with enrollment statistics.
The best easy entry tool I know of for sentiment analysis is called Lingmotif which takes in text files and outputs a nice “dashboard” of visualizations as a web-ready HTML file with related css and javascript that you can drop in your project website directory. You’ll need to register (for free) to be able to import long texts.
I’d also suggest reading this lesson and some of those linked from it to get a better sense of what this methodology does and how it is best interpreted: Sentiment Analysis for Exploratory Data Analysis Lesson at the Programming Historian.