Heigh-Ho, It’s Off to Mine We Go

Heigh-Ho, It’s Off to Mine We Go

Digital Humanities is shaped by the ways we have come to understand and visualize data. Aspects and programming that are used within the digital humanities world are shaping a new way for researchers to present their findings and for audiences to understand and view those findings as well. Text mining is one way that digital humanists have started to gather data in order to present correlations in historical, or even modern artifacts. Mainly used to find and display data about words or phrases in literature, the many text mining programs that researchers use provide researchers with a faster and more effective way to gather and present data.

I will be centering some of my research on art from the American Indian Movement, and how that relates to other activist and Civil Rights groups from around the same time as the founding of AIM. While most of what my research will be centered on won’t be compatible with text-mining, I still might be able to integrate topic modeling and other forms of visualization for my final project.

Ted Underwood, in his blog post, talks about how there are supervised and unsupervised forms of modeling. My project has a clear goal and research topic, which allows for the use of supervised modeling. This supervised modeling can help me to discover patterns that lead to a specific outcome and for a specific purpose. Because I will be focusing my research more on art, I will be able to find correlations within the artifacts that I have gathered; correlations like color, what the art depicts, and similar themes.

While the research that I am focusing around is not ideal for text-mining, there are ways that visualization of text mining and visualization of my research can go hand in hand. Text mining, as Dan Cohen uses it to explore the Victorian Era, focuses on words that can help locate central themes of the time; themes like faith, industry, and science. I can also explore common themes in activist art from the late 1960s through the present to see how the themes have (or haven’t) changed. When it comes to visualizing this data, I will be able to graph these themes, through correlation to each other and/or their growth over time.

One thought on “Heigh-Ho, It’s Off to Mine We Go

  • Very nice Emma!

    One way we might merge your visual interests with text analysis by exploring words associated with the images you have in your imagery. These might be texts included on images, or accompanying metadata. Also, you might include a separate corpus of text documents that you could use to evaluate a theme you explore with your visual sources separately.
    For instance, if you notice that a substantial shift happens in visual/material culture during 1970 (for instance), you might take a corpus of pamphlet or newspaper articles to see if you can identify a concurrent shift in word choice. To build off our discussion yesterday – text analyses on their own don’t necessarily demonstrate causality – neither would visual analyses – but TOGETHER – we have a stronger indication that something significant was changing, which perhaps gets us closer to a causal story.

    By way of textual corpora, remember to check out historical newspapers – Minnesota has a real nice collection of digital newspapers online – https://www.mnhs.org/newspapers/hub you might also sign up for the free trial of newspapers.com…

    The Smithsonian has a nice search portal – https://collections.si.edu/search/index.htm

    Remember to check out the resources on the Reinert History Resources page – http://culibraries.creighton.edu/c.php?g=461228&p=3153777

    Although we won’t cover it in class, you might explore this tool for the images you want to explore – have you been collecting them? http://lab.softwarestudies.com/p/imageplot.html#features1

Leave a Reply

Leave a Reply

Your email address will not be published. Required fields are marked *