Digital Humanities Project Management

By

Cindy Nguyen

·

Week 6 Readings & Responses: Concretizing Text

Final Project Milestone 2: Carving Time & Energy Project Plan – Submit Project Plan as a google doc or google spreadsheet based on this model by February 11 midnight via email to professor.

Note: This week instead of zoom office hours on Wednesday, I will hold office hours right after class on February 12 2-4PM (outdoors, if the weather is good) of Northern Lights. Please come on by, with your group or on your own.

Readings

Complete readings and post your response by Friday February 9. Post a peer comment by Sunday February 11.

  1. Bode, Katherine. “What’s the Matter with Computational Literary Studies.” Critical Inquiry 49, no. 4 (Summer 2023). https://doi.org/10.1086/724943 PDF
  2. (A preview for our in class practice) DSC by Katherine Bowers: Voyant’s Big Day
  3. (The research question iteration) DSC #19 by Quinn Dombrowski and Shelley Staples: Shelley and the Bad Corpus 
  4. Watch: Miriam Posner’s Video “Text Analysis: A Walking Tour of What People are Using in Digital Humanities Right Now”

Preparation for Class 6

  1. For class we will focus on Voyant Tools and this dataset, download the dataset here. Optional, bring your own dataset in .txt file format
  2. Explore and follow along some of the tutorials in Melanie Walsh’s Introduction to Cultural Analytics & Python” – Preview for example the series of lessons on Text Analysis to see the different approaches. Figure out your own relationship to ‘coding’.)

Responses to “Week 6 Readings & Responses: Concretizing Text”

  1. jo alvarado

    I loved reading through The Data-Sitters Club and the ways in which we preserve and tell the narrative of data collection/knowledge production. Over the past few weeks of readings, I’ve been caught up with the emphasis on the humanization of data and how digital humanities as a field centers people. I was especially moved by Katherine Bowers’s memorialization of Stéfan— Voyant, as a tool, reflected his modes of living, researching, and being with people. It is not a mechanism for “cold numbers or a chilly distillation of the written word” but rather an embracing of the endless possibilities of language. Especially being in such a traditional department, I feel the pressure of the institution, of academic labor, and it is encouraging to witness the many scholars who dedicate themselves to ensuring that digital humanities focuses on its people. In a similar vein, Anouk’s addition to Quinn Dombrowski and Shelley Staples’s article that if DH is people, then we must also accept that we are complicated, fallible, and always in flux. I appreciate how transparent these DH scholars are when discussing their hinderances, grounding the field in the reality of the world and concerns that lie outside of academia. This is all to say that I hope to really embody these practices in my work, both in digital humanities and English literary studies, and also both in and out of the institution.

    Like

    Reply
  2. edenwetzel

    It was really interesting getting to read some of the data-sitters club and see the passion Katherine and Quinn and others have for DH. Their narratives of data collection, analysis, and visualization makes the process of learning more about DH more accessible and easy to digest/understand which is something I really appreciate as someone who is newer to the space. It would be nice to see this narrative style used more to help people learn more about DH and other trypes of research in other fields! I think it would help make data collection and analysis easier to comprehend and ultimately help people get more excited about/ confident managing data.

    Like

    Reply
    1. Jo Fobbs (they/she)

      Hi Eden,

      I felt drawn to this week’s readings for similar reasons, and I definitely agree that personable narratives and eye-catching, novel visualizations make DH more accessible! As a visual/kinesthetic learner, seeing the Data-Sitters walk through using Voyant made it feel much more approachable than just reading a how-to article. Somehow, seeing an intriguing visualization and wanting to recreate it made me feel so much less intimidated about working with Voyant for our upcoming class. It tricked my brain into looking further into the process of creating the visualization itself! It’d be interesting to see how similar work could engage other neurodivergent and/or chronically understimulated individuals like myself — not as a ploy to get people to do DH work, but to make the introduction feel much more accommodating and exciting, and make the “busywork” much less arduous.

      Like

      Reply
  3. vmariebarrios

    I found the discussion on the debates over using computational analysis in literary studies very interesting, especially as a literary scholar myself. There seem to be constant debates in the field about whether to modernize, how much to modernize, what is relevant and no longer relevant in literary study, etc. I really enjoyed reading about the work in the Data Sitter’s Club reading, and how you can use text analysis to make insights and visualizations. I am excited to begin using these tools to explore the texts I work with.

    I would be curious to know if other fields of study have similar debates about whethe

    r to use quantification, and computational analyses?

    Like

    Reply
    1. aaronggoodman

      Hello,

      In my undergraduate studies in musicology/ethnomusicology, there were kind of debates surrounding quantification which, like the debate surrounding CLS in Bode’s piece, were somewhat philosophically charged. While these debates concern methods that in some way quantify music, the methods in question are far simpler and less constitutive of “computation” than the examples we are looking at involving literature and other text. In particular these ethnomusicology debates may concern whether marginal differences in pitch constitute a unique music or musical culture. One of my professors was a scholar who got nitpicky with scales, for example, and he conducted a lot of research which systematically accounted for regional variations in musical expression via mode or scale. While he was an expert and he knew extremely slight differences in how various cultural groups performed certain modes, he never divorced his mathematical measurements of pitch from the sociocultural meaning that the note/sound held in the scale or song. Most of our seminar time was spent addressing this balance in quantifying music.

      Like

      Reply
      1. edenwetzel

        This is a great question! I think often in the sciences there is a push towards computational analysis and a pressure to stray from more qualitative, exploratory anaylses. The main debate I feel tends to stem from perceptions that computational analysis is more valuable than the qualitative research. However, both are important parts of the story of the data being analyzed and I think both methods should be used more in conjunction with one another in the sciences.

        Like

        Reply
  4. collinmoat

    Takeaway: I enjoyed Miriam Posner’s video and DSC articles this week. It was nice to get a general overview of tools of textual analysis and then a run down of how to use Voyant. Quinn’s tutorial of digging into a corpus was an eye-opening exercise. I liked that she demonstrated how a simple question can branch into many just by looking at the data a different way. It’s a great reminder that as we ask new questions, we always need to adjust and refine our understanding of the corpus, especially its limitations. Also, something that Miriam said really stuck out to me: it is when you layer various tools that you start to uncover meaning in the text. I will keep this in mind when working on our project’s corpus in the coming weeks and try to explore how various combinations of tools can illuminate different facets of the corpus and spark new questions. 

    Clarification Question: As Quinn shows, the exploratory and research phase can splinter into a thousand questions worth asking, but when do we stop asking questions in a rabbit-hole kind of way and start giving answers? Maybe when we run out of energy or some deadline is upon us?

    Discussion Question: In response to the importance of frequency in the tools of textual analysis, I’ve been thinking about the place of the unique in all of this. I study Homeric poetry, and hapax legomena “things spoken once” often are subjects of scholarly buzz. They are sometimes sites of interpretative difficulty, indicators of thematic importance, or glimpses into contexts tangential to the one in focus. What do we gain from starting from the end of the frequency list and looking at unique words in a corpus? How can we balance our interpretation of the often spoken with the rarely said? Have there been times in your own reading or research when infrequency gave you more to think about than frequency?

    Like

    Reply
    1. vmariebarrios

      I think this is a great question! There are oftentimes in literature when the rarity of a word/phrase is what makes it important. I can also think of times where there is a “trend” in frequency only for it to be disrupted later by a different word, and that difference is important. Makes me think of how we can use these tools flexibly to understand these infrequencies!

      Liked by 1 person

      Reply
    2. jo alvarado

      I love your question about frequency vs. infrequency and the possible analyses we can construct when investigating the latter. I study contemporary Asian American poetry, and I’ve also been interested in absences, lack, and the things left unsaid. Especially as someone who centers inscrutability and illegibility, the focus on what isn’t there, on refusal as a practice, is something endlessly captivating.

      Like

      Reply
    3. sparksdv

      I wonder that same thing about exploratory research. Do we set boundaries with what we explore or would that directly hinder the process

      Like

      Reply
  5. aaronggoodman

    I really appreciated Miriam Posner’s “tour” video because the various examples of methods in use efficiently illustrated the pros and cons of a particular method. Vector analysis in particular seems both strenuous and massively powerful; I would be curious to see what kinds of vector analysis have been attempted for musical texts, i.e. lyrics or other sung poems.

    Bode’s piece on “Computational Literary Studies” was quite a head scratcher but I appreciated the depth with which she covered stagnating debates in CLS. It seems to me that regardless of how one scholar aligns themselves philosophically, they won’t be convinced of CLS or other computational methods’ efficacies until they see it practiced on data they care about; maybe I’m being too cynical. I think the care and transparency with which DH scholars employ their methods is evidence supporting pushes for further experimentation with computation in fields like literary studies. But many people may not believe that a computer can do XYZ until they see that someone else made it do that.

    I’m curious if others have had any personal experience with teachers, advisors, or peers who are particularly outspoken in support or disapproval of computation in your field.

    Like

    Reply
  6. Melinda Linyu Xiang

    This week’s readings provide a more detailed and practical understanding of context analysis, introducing theories, critiques, and tools. I particularly appreciate the DSC articles, as they offer step-by-step instructions on how to analyze a text. The accompanying video serves as a general introduction to the specific techniques and software involved in text analysis.

    My initial exposure to text analysis was during my undergraduate studies, where I utilized Natural Language Processing (NLP). The NLP tool I employed focused primarily on interpreting the emotions or feelings generated from the text, along with determining the negative or positive tendencies. This, in my perspective, heavily relied on manually preset corpora related to various emotions. However, the lingering question of “How do you know when you’ve got enough?” weighs on my mind.

    Simultaneously, the phrase “all models are wrong” resonates with me. It underscores the inherent limitation that models cannot encompass all the factors we encounter or address every research question. In analytics, the objective is to exclude less relevant elements and keep the model focused on the main road. Similarly, when creating corpora or conducting text analysis, it’s essential to recognize that we can only include what matters most to the research question. Acknowledging the necessity to exclude certain relevant but less significant elements is crucial.

    Moreover, understanding that human emotions are delicate and complex reinforces the idea that even the most sophisticated machine learning cannot yield the most satisfactory results. In conclusion, the journey of text analysis requires a mindful selection of relevant factors and a humble acknowledgment of the inherent limitations in capturing the intricacies of human emotions.

    Like

    Reply
    1. esperanzabey

      Hi Melinda,

      I love your final point in addressing how human emotions are delicate and the limitations of machine learning. I think especially in the age of all these AI initiatives, it’s important to see value in recognizing human emotion rather than teaching machines to recognize human emotions. Great point!

      Like

      Reply
  7. Jo Fobbs (they/she)

    I really enjoyed the Data-Sitters Club reading and Dr. Posner’s video on some of the current methods of text analysis, mainly due to their focus on visualization (pretty colors are very stimulating). I was very intrigued by the collocates graph, which, although used in the context of slang, made me think of its potential uses for flagging subtly discriminatory terms in texts that often go unchecked. This brought to mind the cover of Dr. Safiya Noble’s book, Algorithms of Oppression, where typing in the phrase “why are Black women so” into Google resulted in a series of negative search suggestions. I’d be interested to see a collocates graph of the word “strong” in a corpus of medical texts about Black women. Something like this, if not already done in another project, could be very useful for exposing the problematic nature of not just the words themselves, but the context they’re situated in. Though, this brings me to question — what are the limits to the scope of a corpus, and how much variation in topic is allowed before it would be considered too broad?

    I also really loved the Signs @ 40 models, particularly the “topics over time” visualization. If you take away the labels and crop portions of this visualization, you could get several unique pieces of abstract art. Who, other than a DH scholar or artist with a sharp eye, would know that this is data? It’d be beautiful to see such visualizations be used in an artwork, but that begs the question — is that real “art”, and moreover is it still really data?

    Like

    Reply
  8. craigdavidsmith

    I found Miriam Posner’s “Text Analysis: A Walking Tour” to be quite informative. It was useful to see practical examples of a various text analysis methodologies and the strengths and weaknesses of each. Although I have seen examples of word clouds before, this tour has encouraged me to use them more frequently. I think they can be a very effective tool as they are so easily understood by all viewers without the need for any further exposition. I feel that concordance tools could be helpful to my research. It was also nice get an introduction to the “Data Sitters Club” and Voyant in Katherine Bowers post.

    Like

    Reply
  9. sparksdv

    The reading The data-sitters club and Dr. Posner’s video were very enjoyable and understanding the current methods and ways we can visualize text analysis. It was great to see the pros and cons along with the practical examples of these methodologies.  I enjoyed visual nature of bring a paper to life as another way of reading it. I wonder how CLS would be incorporated in looking aliment over various ancient text or poems as they can be difficult to interpret in understanding culturally significant themes. I have been really struggling trying to understand how this could be useful within archaeology to make our research more expansive. What are some ways you have found frequency useful in research?

    Like

    Reply

Leave a comment

Design a site like this with WordPress.com
Get started