Text Analysis: Religion and Slavery – Page 2

These trends can be greater analyzed by conducting a closer reading of the documents. For example, slave and religious terminology may have been avoided being spoken of together throughout the mid 18th century, because of the American Revolution and that ideas of freedom and liberty were not offered to the enslaved. It is also possible that a more even distribution of religious and slave terminology occurs throughout the mid 19th century, because of the American civil war, its precursors and aftermath.

The next style of visualization that further supports the trends/categories mentioned above is a Stream Graph.

Analyzing the width of each colored category within the stream showcases the frequency of words within each document. The documents are also listed chronologically. Stream graphs are better at showing the change in frequency of word use over time. One issue with using this graph over Bubblelines is that one cannot visualize where words are used within a specific document. For example, the ability to see that slave and religious terminology are mentioned at opposite ends of the document from 1734 is not available when using a stream graph

A third style of visualization that can be utilized to analyze the use of religious and slave terminology within these documents is a MicroSearch graph.

Like Bubblelines, a MicroSearch graph allows a researcher to see where terms are dispersed throughout a document. A major problem is that they are all displayed as the same color. This makes it possible for a researcher to skew their data. For example, the document 1810-whitegeo.txt shows up on the graph highlighted more than any other document. If the terms christian and god are eliminated from the graph, then it becomes one of the least highlighted documents. Thus, a researcher must keep in mind the differentiating factors of a MicroSearch graph as terms are added, and not solely the final graph.

One point that a MicroSearch graph emphasizes that Bubblelines does not is the size of each document relative to all of the other sources. In doing a deeper reading of the documents, one should always be aware of the size and style of the primary sources that they are using. For example, trends might develop between the difference in content and meaning of a short letter compared to a longer one, or a pamphlet compared to a book.

Ultimately, of the three styles of visualizations above, Bubblelines is the best. Stream Graphs allow a researcher to visualize differences between terms, but not where the terms are situated within a document. MicroSearch is the opposite. It allows you to see where terms show up in a document, but there is no distinction from one term to the next. Bubblelines permit the user to do both.

Pages: 1 2