Early Amherst Perspectives

After many project proposals, methodology workshops, and blog posts, I think I am ready to start crunching out a final research project website deliverable. Design is an iterative process. Although our brainstorming has somewhat narrowed our focus and pointed to where we need to invest our time for the remaining three weeks (time flies when you’re having fun!), I feel the need to start molding the clay of research accrued over the past few weeks. The challenge, now, is to decide what what sculpture we as interns want to (and can) make in the remaining time (some time has been invested into brainstorming the possibilities), where the online piece will be hosted, and what tools we will call on.

One mini-project that has captured my interest is the “Amherst___through the lens of___” project. Potential fill-ins for the blanks could lead to each of our different projects, which focus on early Amherst perspectives: social networks, the early College library collection, academics as viewed through course catalogs, and architecture. What is interesting about this proposal is its potential to draw the site user into learning about early Amherst in a playful, interactive way. In giving the user autonomy to select their path through the site according to what lens interests them the most (be it Amherst faculty, a specific student or society, or time period, or even a specific building) the site will keep the user engaged as they leap from one section to another. A challenge, however, would be to make the different sections overlap enough to make a scholarly argument about early Amherst, as Este suggested in our team meeting.

As critical as it is to produce an aesthetically pleasing, fun website, it is equally important to see this assignment as a scholarly research project. While we have been absorbing information in preparation for our final product, we should keep in mind that the site must balance between disseminating information and advancing scholarly work on the data available on early Amherst. I can’t believe I’m beginning to sound like my thesis advisor!

So, that said, I shall start site construction this week. I will get frustrated with the tools I have and my shortcomings in using them. I will fail to make the visuals match my vision for the site. I will miss some important data. But I will learn. I will get better at using the tools, and I will research more to find missing pieces to the puzzle. In the words of Amherst alumn William Hastie (1925), “Achievement can be all the more satisfying because of obstacles surmounted.”

Without further delay, may the site construction games begin!

A Method to (DH) Madness

I must admit: after the first week of my Digital Scholars internship, I thought the task of researching the early college history in the span of two months was insurmountable. Among many linear feet of manuscripts, countless volumes of publications, articles, and journals, apparently, lies new insights into the early college history that I must dig out. This task beats finding a needle in a haystack for difficulty, I thought. I equated it to finding a silver one hidden among a needle-stack in fifty shades of gray, all within a limited time frame – nearly impossible. After an additional week of methodology workshops, however, I found my concerns abated.

This week focused our attention as interns on text analysis techniques: Google Ngrams, Voyant, Lexos, and topic modeling. In addition to learning how to distill large volumes of text, I picked up a few new words that allow for better understanding of the hermeneutics of my corpus (I may need practise at using these new words though). I have come to understand the methodologies applied to Digital Humanities in a practical way (as is natural for my architecture background). Like a fulcrum, text analysis tools do not change the load of information to be lifted from the Archives and Special Collections (pun always intended). Rather, the tools allow for more output for the effort placed into analyzing large volumes of text in a limited span of time.

I will not go into the details of the features of each of the tools we learned mostly because I am yet to fully grasp each of them, and partly because they each achieve similar outcomes: to translate texts into graphic information. Text analysis is a neat art! As a visual learner I appreciate how, for example, a phrase or argument can be traced in a body of text, or across different texts that may or may not be explicitly related. This is valuable in our quest as interns to acquire new insights into the old material available in the Archives and Special Collections.

The text analysis workshops have reshaped my approach to my project for the internship. Rather than exclusively focus on using visual material such as photographs and architectural drawings to understand early Amherst College architecture, I will be analyzing college publications and journals from between 1821 – 1861 to compliment my findings thus far. Previously, I was overwhelemed by the quantity of the material available for the scope of our research. Now, given additional time-saving tools, I am ready to begin analysis of texts that point to the rich early college architecture.

I cannot say that I have mastered many of the new research tools we have been taught. Nonetheless, I feel more confident that the task before us is possible given our awareness of more efficient ways to climb the mountain of material before us. It seems, afterall, there is a method to this madness.

Can We Guess What Your Learning Preferences Are With This One Question??

One of my favorite little things about this internship so far has been seeing how everyone in the group approaches our weekly blog posts so differently. Although we all are given the same prompt and have the same tools available for answering the prompt, from the beginning, each individual has made their blog post unique, distinguishing themselves with stylistic choices like bullet-pointed narrative, allusive featured images, or a first-person reflective voice. At first, I was surprised at the dramatic differences between the look and feel of all of our blog posts, but after a couple of weeks with the team, it makes sense. After we had a workshop that helped us determine our learning style preferences using the Kolb learning style inventory, it was apparent that the four of us all had very different preferred learning styles. Even where a couple of us elapsed in preferences, the determined way that we expressed those preferences was different.

The interesting thing about the Kolb inventory is that it presents the learning styles not as discrete, but all making up part of a cycle of learning that we all experience (albeit in different ways and at different paces).

So while my preferences fall in the “Assimilating” category (which prioritizes observation and thinking before any of the other steps), at some point I will have to move on to the next quadrant of thinking and doing, testing out the ideas that I’ve been conceptualizing.

This is the point I feel as though I’ve reached in my section of the research project. I spent last week collecting a large amount of data on when five of Edward Hitchcock’s most important works have been cited, according to Google Scholar. I have a nice big Google Sheet with individualized tabs and lots of data. The question is, where do I go from here? The goal with this branch of the project is to map the network of Hitchcock’s scholarly influence after his death, but given the diversity of data I’ve collected, this could present itself in a variety of different ways. Do I map exclusively the numbers of citations and co-citations? Should frequently appearing authors or journals be connected in some way? Does that matter to us? (It might show that what appears to be a very widely spread network is just a network that is very insular, but active.) What about mapping with an emphasis on time and place? These are elements I’ve been very interested in recording from the beginning, and I think they also have some relevance with the data Seanna’s collecting, at least on a macro, if not quite micro level.

Screen Shot 2015-07-27 at 2.01.15 PM

The time period of Seanna’s project is focused on the years of Hitchcock’s life, when he was actively publishing. Additionally, the data she’s been collecting comes from sources that were relatively geographically close to Hitchcock as well; the furthest source I’ve seen so far recorded is London. At first I was bothered by the lack of overlap in our data, but now looking at the scopes of our respective projects side-by-side, it seems like they’re natural continuations of one another that could easily segue back and forth. I’m envisioning two separate networks with different data focuses (foci?), but when you “zoom” in or out of one you reach the other. In the end, they’re both trying to measure Hitchcock’s relevance, influence, and effect on his intellectual peers.

This is my dream architecture, the two networks represented visually with zooming capabilities and detailed nodes. However, this is my imagined approach, one that is very visual and very attached to the network image that’s been fascinating me from the beginning. Like with the blog post, I’m approaching this project with a preconceived set of expectations for what it should look like. When I find myself hesitating about the direction to take my visualization (and especially in what data to include), I think that one way of moving forward could be turning to the team to see how they would approach the visualization. If we had more time (always, always, if we had more time), I would love to take a day or half a day to have sandbox time with the data and Tableau. During our Tableau workshop, we played around with some of the sample data sets provided on the site, and it was fascinating to see what everyone chose, and then how they chose to visualize it with the software. I think it would be a great brainstorming experiment to give everyone my data and tell them to mess around in Tableau and present it however they thought was best. If we want to get more complicated, I could ask them to arrange the data with an emphasis on a certain category or using a certain type of visualization, to compare and contrast how they (and a potential later viewer) might be interested in looking at the data.

I’ve been trolling around the Tableau website looking at the sample visualizations they have as a substitute for the above experiment. They really showcase the variety of options you have in Tableau for displaying data, which reassures me that there’s no set template for visualizing data in a certain way. I’m hoping that through some more experimentation, ideally with the input and aid of my fellow team members, we’ll be able to find a method of visualization that best showcases the network data.

 

Making knowledge of data: from the analog to the digital in humanities research

What constitutes digital humanities?

It is a question that eludes even the professionals and scholars of the field—let alone me, a humble student intern. There are many answers to the question, most of which can be categorized into three basic camps of thought: the crusaders, the conservatives, and the cynics. The first camp consists of those who believe that DH has the potential to disrupt and transform the world of information and knowledge. They are optimistic if not utopian. It is the realists who comprise the second camp. They recognize DH as a set of new digital tools that can augment more traditional humanities scholarship. If the crusaders explicate the reaches of DH, then the conservatives delineate its limits. Finally, there are the cynics. This censorious bunch believes DH to be the swan song of humanities itself, a last ditch effort made by increasingly defunded (read: irrelevant in today’s market society) humanities departments across the United States and the world at large.

Whichever camp one agrees and aligns herself with, it is relatively noninflammatory and perhaps agreeable to say that the trend towards the digital in the humanities bespeaks a wider trend toward the digital in our culture. In fact, the insight is..well, unremarkable.

We live in the era of big data, of amassing Brobdingnagian inventories of information so that they may be mined for specific purposes of either a commercial or educational nature, mostly. Statistical analysis and summarization is a useful skill to have nowadays, and the salaries for entry-level jobs in the tech world help support such a claim. What then is digital humanities without data? Documents such as articles, books, certificates, citations, film, illustrations, letters, photographs, receipts, and many more objects crowding archives everywhere all contain data within them. Traditional scholars make use of that data, or information; they examine it, unpack it, and assemble it so as to produce new knowledge—at least, that is the ideal of the métier.

It is that process of scholarship, or rather, data analytics I hope to replicate this summer with my fellow digital scholarship interns as we work with the digital collection of the Edward and Orra Hitchcock Papers. Some of the questions such an approach raises includes: What data lies dormant in the collection? How can it be surfaced and organized? What can we say about the data? What does the data reveal? How does it enrich our understanding of the lives of Edward and Orra Hitchcock? How do we impart our findings to others in an accessible and engaging way?

My aim is to have stimulating, thoughtful  answers to these and other questions that may surface along the way—answers that may help begin to illuminate the future direction of our interaction with the past through new means of technology. This is the start of that endeavor. Where we end up remains to be seen.

The Strengths and Limits of “Google Squids”

The beginning of this week, I headed to the nearby University of Massachusetts Amherst for a NERCOMP conference! I know it sounds like NERD-COMP! To be fair, we were all kind of geeking out about data and other techie stuff all day though…Anyway, during the data visualization workshop that we attended, I spent a good chunk of the exploratory “sandbox” period evaluating Google Fusion tables. What’s really cool about Google Fusion Tables is that you can take data from spreadsheet columns or CSV files, and create quick visualizations. In particular, I tackled the Network Graph feature. According to Google, “this type of visualization illuminates relationships between entities. Entities are displayed as round nodes and lines show the relationships between them.”

Continue reading The Strengths and Limits of “Google Squids”

Tech + Text

 

What makes a DH/DS project work or splutter out depends partly on the wedding of digital tool and project materials. With the wrong combination, the whole project can go awry. This week we took the time to consider how the KWE Native American book collection might cooperate with one of the tools we’ve “sandboxed” to get a feel for.

  • ArcGIS. After completing a four-day, twelve-hour workshop in ArcGIS, we got a feel for the capabilities of importing census data, using different map projections, and layering on features like rivers or elevation data. In theory, this could provide a way to look at the KWE, perhaps using locales mentioned in the texts or mapping out the publishing houses over the decades. Continue reading Tech + Text