Data Mining

Hallmark Christmas and Victorian Tattoos

I have a confession to make.  This year has been difficult and to make it through the holiday season, I have found myself watching a lot of light-hearted holiday fair.  I’m embarrassed to admit the majority of which is from the Hallmark collection.  Why am I writing about this for a digital humanities assignment?  Well, when playing around with the Google Books NGram Viewer, and having just recently finished the Netflix holiday series Dash and Lily (don’t judge!). I began typing common holiday phrases into the search bar, such as Christmas eve, Christmas morning, holiday movie, marzipan (I love the British Baking Show), and even Hallmark Christmas.  The results showed a steady increase of popularity in the late 1990s and early 2000s, peaking around 2019.  My hypothesis is that the trajectory of the phrases matches the history of Hallmark holiday programs, which aired its first original holiday movie in 2000 and expanded the programming in 2009 with dozens of original movies being aired during the “Countdown to Christmas.”  Today, the term Hallmark Christmas has multiple meanings and plays a significant role in popular culture during the holiday season.

 

Google NGram Search: Hallmark Christmas
Google NGram Search: Christmas Movie

My sojourn into the history of campy and plot-predictable Christmas movie history is a simplistic example of how data mining can provide insight into digital humanities research. For example, Professor Robert Shoemaker and Dr. Zoe Alker’s research into Victorian era convict tattoos challenged previous beliefs that tattooing was an expression of criminal fraternity.  By utilizing data mining techniques centered on the Old Bailey convict records housed in the digital panopticon – 75,688 tattoos on 58,002 convicts in Britain and Australia from 1793 to 1925 – the researchers were able to show that tattooing was in fact an acceptable practice of working people during the time period.  The use of data mining allowed the researchers to view the tattoo records with a new light while discovering associations and patterns not seen before.  They looked into already established datasets, created visualizations (example chart below), and identified the types and locations of popular and recurring tattoo themes which opened up a window into the lives of ordinary working-class people in Victorian London.  Similarly, Ruby Mendehall’s research on the historical experiences of black women analyzed 20,000 documents from the HathiTrust and JSTOR using high-power computing.  The researchers knew that the documents contained information on black women but they needed to create methods for extracting the information. In the end, they used data visualization techniques and topic modeling to make senses of the findings.

Digital Panopticon: Female Tattoo Subjects 1830-1925

Using computing tools and machine learning to enhance humanities focused research is an interesting concept. It is clear that data mining practices can speed up research tasks while uncovering hidden and interesting patterns in established or overlook data. As an outsider reviewing several of the datasets available I could see how specific questions could remain unanswered without the use of high-powered analysis techniques and methodology.  And, because of data mining research I now know that Buffalo Bill was a popular tattoo in Victorian London and Hallmark Christmas movies remain a popular escape for troubling times.

3 Replies to “Data Mining”

  1. Melissa,

    I love the Christmas theme of your blog post, way to get into the holiday spirit! It’s very lovely. I was personally fascinated by your search. When you searched Christmas eve, Christmas morning, holiday movie, marzipan and even Hallmark Christmas, what other patterns or things did you see? That’s really cool that the results showed” a steady increase of popularity in the late 1990s and early 2000s.”

Leave a Reply

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