For textual analysis, we analyzed the frequency of the words selected to appear on the carleton.edu homepage and studied the trends in word choice. Our goal was to see why the text changed over time and how that affected the message that Carleton was trying to get across on its homepage.
Accessing the Data
To get the necessary webpages we needed for the project, we used the Wayback Machine. Each year, we selected the final snapshot of the year. We used Voyant – a web-based reading and analysis environment for digital texts – in order to “strip” down the webpages and analyze the text specifically. With just a link to the website (from the Wayback Machine), Voyant would analyze the text and format visual representations of them. The only issue we had was that, with just the link, text that wasn’t on the forefront of the webpages would also be included in the analysis. Therefore, we used a filtering tool included in Voyant so that it would only analyze the text on the webpages that readers could see.
We used Voyant in order to find the word frequency of each year, every 5 years, each decade, and overall.
Analyzing the Data
To analyze the data further, we created spreadsheets and wrote down the top 15 words of each year. For every 5 years, decade, and overall, it was the top 20 words. We would use this data for analyzing the trends.
Below are word clouds that illustrate the frequency of words every 5 years. Click on the links below for more in-depth analysis:
