Upload
kory-powers
View
216
Download
0
Tags:
Embed Size (px)
Citation preview
1
Textual Portraits: Using Word Clouds as an Analysis Tool for Digitized Texts
Shonn Haren
Wichita State University Libraries
February 28, 2015
2
Visualizing the Text
“…often the most effective way to describe, explore and summarize a set of numbers-even a very large set-is to look at pictures of those numbers.”
– Edward Tufte, The Visual Display of Quantitative Information
3
Word Clouds for Text Search
• Google Books– https://books.google.com– Accessible from the books’
record page– Click on a word to see the places
where it occurs within the text.
4
Word Clouds for Textual Analysis
• Brad Borevitz’ State of the Union Website– http://
stateoftheunion.onetwothree.net/index.shtml
– Compare the clouds of Speeches from Washington to Obama
– Track the use of terms over time
5
What are Word Clouds?
• Machine-readable text is broken down into its component words
• Common words (a, an, the, etc…) are filtered out
• Remaining words are assigned values based on frequency of use
• Words arranged randomly in a cloud, with their size determined by their frequency of use:
6
The Bill of Rights
7
The Gettysburg Address
8
Comparing Word Clouds
Topeka Constitution Lecompton Constitution
9
Comparing Word Clouds
Topeka Constitution Lecompton Constitution
11
Benefits and Drawbacks
• Benefits– Wordle.net is free to use– Shallow Learning Curve– Images produced can be
used for outreach too
• Drawbacks– Text must be machine-
readable– Synonyms aren’t
eliminated– Archaic
texts/translations have uncommon common words that must be manually removed
12
In Conclusion…
• Word Clouds don’t replace traditional textual analysis or reading
• A first glance tool for comparing unfamiliar texts with minimal metadata
13
Any Questions?