For my final project, I would like to explore the possibility of proposing a collaboration – with the current DH project Moveable: Narratives of Recovery and Place (based out of Marshall University and West Virginia) to create additional features/layers for the current iteration of their story map and archive. Moveable, a live DH interactive mapping project that chronicles personal stories of addiction and recovery in Appalachia “and beyond,” is an incredible collection of text-based first-person accounts of recovery and addiction. As powerful as the project is, it seems to stop at Appalachia, and perhaps falls short in appealing to users who want to browse the material in more flexible ways to better understand the relationship between time, place, addiction and recovery. The project is live, and currently being updated, so I am interested in exploring how to make these stories even more accessible and fluid — How can we move beyond lengthy text-based storytelling in favor of a multimedia approach without diminishing the power and sensitive nature of the project’s stories and people? How can we continue to modernize a DH tool to appeal to users with increasingly shorter attention spans?
The Movable Project: a platform for people in Appalachia and beyond to share, highlight, and document stories of recovery.
Features to propose and consider:
Timeline – to help map and chronicle the difficult journey of recovery in time – one that often looks more like a circle, starting, stopping and restarting
Additional Map Layer – To toggle between traditional views like city/state/country to more personal data points like breaking points, rock bottoms, moments of hope, feelings of “home,” journeys made from “beginning” to “end” and in between
I would love feedback on whether or not one can propose a collaboration or addendum to a current DH project, and how to go about creating such a proposal without necessarily critiquing such a meaningful and nuanced digital archive. I’m leaning towards focusing on one feature over the other, but not sure whether that is a worthwhile addendum to a pre-existing project, or something that should be explored in a different project altogether. I love the idea of collaborating and building off stories already shared, but not sure about the ethical considerations of expanding work already collected and created by and for a particular tool.
As someone who works in the restaurant industry, I am always thinking about food and dining. The COVID Pandemic had, and continues to have, a major impact on the industry, so I was excited to dig deeper into its effects, particularly in New York City, through this project.
Coming out every week, The New Yorker “Tables for Two” restaurant reviews have remained a staple in New York City food writing, even through the pandemic. While it might not be the most robust writing on the state of food more generally, I thought it would be a good place to start for analyzing trends in dining. It also has a dramatic impact on restaurants, and is sometimes responsible for huge booms in visitation.
As I’ve shared in class, I have trouble coming up with research questions, but always know the sorts of topics I’m interested in. I shared this with Filipa and she noted that sometimes with text mining, it’s best to simply upload the corpus in question and see what comes up. I chose to work with all of the “Tables for Two” reviews of 2019 and 2020 in order to have a bigger corpus to work from, and to be able to compare the before and after effects of the pandemic more specifically. I knew I wanted to focus more deeply on the corpus rather than learning a new software/tool for the assignment, so I choose Voyant for its ease of uploading and working with texts.
Starting with Voyant
Getting started, I tried simply uploading the links to all of the reviews, but came across a mess of words. Each New Yorker digital review contains links to other articles published in said edition, along with dozens of links to various New Yorker digital features and common website lingo. After inputting the webpages for analysis, the most common words that appeared in the word cloud were not very helpful:
While words from the site and related articles could certainly prove interesting to explore (and definitely heightened the use of the word pandemic), it proved too time consuming to try and remove each repetitive word using Voyant’s “Define” option. I also found that said option was not very successful, and often kept words and variants of words I hoped to remove in the cloud. In turn, I went the old fashioned way of cleaning my data, and copied and pasted only the text of each article into a Word Document. Of course, this still required some cleaning, so after some more massaging of removing words like “new” “restaurant” “food” and “yorker,” I got started exploring my texts.
(As a quick aside, I took some time to really think about what it means to remove words from a textual analysis. Every text comes with context, and it felt a little like cheating to be removing the reviews from their origin point – especially if I wanted to compare the state of dining to the rest of the world in 2020. That said, it ultimately felt like doing so would create a different project altogether — or maybe, could be something to focus on for my final project.
Before getting started, I also wanted to learn a bit more about the different features that Voyant offered as a new user, so I watched a few really helpful YouTube videos. I posted them below for anyone interested:
Most commonly used words in 2020 “Tables for Two” reviews
Most commonly used words in 2019 “Tables for Two” reviews
I wanted to start easy – what words have become more common in restaurant reviews this year than in the previous year? Unsurprisingly, words like “pandemic,” “home,” “frozen,” “closed,” “cooking,” and “takeout” came to the top in 2020. Alternatively, “pandemic” “frozen” and “closed” were not featured in any 2019 reviews, and “home” was only used in relation to discussing a chef or restauranteur’s “hometown”.
The popularity of the word “chicken” in 2020 was a surprise, so I did a bit more research, and came upon articles on the chicken shortage in America during COVID-19.Here, I was able to see popularity of an item during COVID that correlated to a larger food shortage in the country. Interesting! The popularity of the word “people” in 2020 also caught my eye, so I looked to compare its use with 2019 using the “Context” feature:
20202019
I’m not sure if you would call this a “sentiment” analysis, but you can certainly see the growth in relating the word “people” to more complex issues in 2020. In other words, concepts around “people” in dining and restaurants in 2020 has expanded beyond the world of food in 2020 into conversations of equity and need. Seemed like a plus!
That said, I was surprised to see the lack of conversations on racial equity in particular, given the BLM protests in 2020 that sparked discussions of white supremacy in the industry. Here is where I wished I would have done things differently, but kept my mistake for the sake of learning:
I was hoping to see if there was a trend in speaking about black-owned restaurants during the BLM protests that did not continue into the rest of 2020. As we’ve discussed in class, that summer often resulted in lip service to black populations, rather than actual moves towards equity. Since I did not categorize my reviews by month (which would have required separate Word Documents per month), I was only able to analyze trends as a whole in 2020. This made me realize that Voyant is really a tool used best when comparing different texts as whole units rather than comparing a single text as a unit. Since I didn’t go the time route, I looked at how the word “black” was used in the reviews in 2020 vs. 2019:
2020
2019
The screenshots are unclear since Voyant could not seem to finish loading this analysis, but it shows that in 2020 the word “black” was used with the word “entrepreneur” and “lives” matter” vs. 2019 with “tart” “avocado” and “pepper”. Of course, these results don’t look so good for The New Yorker, and I’m not surprised.
As a final exploration, I went to another corpus, Whetsone Magazine, and their 2020 digital articles on food during the pandemic. Whetsone Magazine is a black-led publication on food by Stephen Satterfield. Whetsone’s 2020 article word cloud did not even contain words like “pandemic” or “takeout” but rather words like “family” “father” “women” and “love”. This reminded me of conversations that we’ve also had in class around what types of content is shared by communities facing trauma, and where words like “joy” and “love” fit in. That said, of course it’s important to also mention that Whetsone’s 2020 articles range in content other than just restaurant reviews, but it shows a different sort of focus on eating during a global crisis.
Whetstone Magazine’s 2020 Articles
Where to go from here
Overall, I struggled with this project in that I felt the tool really just helped to prove assumptions I had about texts, rather than surprise me with new learnings. Of course there is always a use case for proving yourself right! Next time I use a tool like Voyant, I would try to focus on further categorizing texts before I upload them for analysis by things liketime, genre or author, in order to get some more nuanced readings of subjects through comparison.
If I were to continue this project for my final project in the course, I would be interested in comparing reviews from either different publications, cities, or topics rather than years…asking research questions like:
Which cities saw the biggest changes in approaches to dining out in 2020?
What publications most holistically reviewed the impacts of the BLM movement on restaurant equity in 2020?
How did changes in dining out compare to other service industries like theatre, film or hospitality more generally?
This week, I was struck by Ryan Cordell’s piece “How Not to Teach Digital Humanities” given its approach to questioning basically the structure of our own DH Intro course. I have been finding some DH readings difficult, especially those that play “inside baseball” and position entry-level scholars to save a field they have yet to enter (a trap that Cordell falls into himself). I have been asking myself, outside from the academic context, why is DH important?
Risam offers a few ideas, particularly when around with DH x Postcolonial education:
Understand the politics of knowledge production and how print, and now digital, cultures marginalize communities outside the Global North
Become critical tech consumers, which can help transform into skepticism into action
Engage with the “core concepts that undergird modernity and alternative perspectives on community formation by considering instantiations in the digital cultural record”
In terms of a pedagogical approach, Cordell gives a few reasons as to why readers like me tend to not be engaged; a lack of attention to case studies in class, teaching DH to DH (rather than to a specific discipline that could benefit from digital tools/theories), and general undergraduate technological skepticism. All of these resonated, and made me question: Where was the digital in my undergraduate liberal arts degree? Should I now just be studying Art or History and use digital tools rather than studying DH itself? Cordell lends a few suggestions that I particularly agreed with (and wish had been shared with me, especially at the undergraduate level):
“You do not need an entire DH curriculum, or even a designated DH course, to introduce substantial digital pedagogy into your classes: Teach distant reading alongside close reading and do not worry about proving how revolutionary the former is.”
“Teach new technological tools as a development in the range of human technology: By contextualizing our moment of digital remediation historically, as but the latest phase in a long history of textual reinvention, help students understand why assignments ask them to experiment across modalities [and] consider the medium as well as the message of their own research and arguments.“
However, beyond pedagogical suggestions Cordell leaves us at a dead end. He writes that once the DH mania is over, there must be “more productive rapprochement with the larger humanities fields.” Where is that? Is the DH mania over? If so, what are these rapprochements, and how can I, as a new student, start from these rapprochements rather than helping to define a field that may one day eclipse itself? I tend to think that this answer lies in practice, and so I look forward to continuing these thoughts and questions as I work on my text analysis project, and into my final work for the course.
This has been my favorite week of readings thus far, lending a new way of engaging with my mapping project investigating food supply chains. In addition, as an interdisciplinary learner, the concept of studying the relationships between networks (visible and invisible) is something very intriguing to me. Novinskie’s work on creating a praxis of “care” particularly struck me as it afforded an extra layer of meaning and truth-seeking to focusing on even the smallest of sites/topics within larger networks to expose their invisibility.
That said, I was also moved by Jackson’s piece “Rethinking Repair” and how to flip my project in, say, focusing on “breakdowns” and “errors” in food supply chains as the impetus for study. This might be in the form of tracking stale food or evaluating waste. For me, the concept of studying sites of repair, and in a sense, error, is also something I’ve come across before in thinking of all of the various types of foods and common ingredients today that have come as a result of an “accident.” Could I map those interactions, sometimes a cultural misunderstanding, in working through their “breakdowns”?
On the other hand, I could take Gil’s approach in illustrating a “technology of disobedience,” and showcasing how/where we make use of foods and ingredients for other purposes – which reminded my of all the things my Italian grandmother does in the kitchen.
I could also think further about Posner’s article in thinking just how disruptive a trace of an otherwise invisible food supply chain is for a capitalist economy based on scale. Reading Posner’s work, my trouble finding sources for many of the foods in my home felt right on target, and it got me thinking about the various food businesses who might already be employing a sort of blockchain technology – including Nestle and Walmart:
Of course, there are quite a few benefits if companies were to do this: fresher food, waste reduction, guaranteed “uniqueness” of a product, improved food safety and more. That said, the probability of adoption feels low given the raise in prices it would spark for both consumer and producer, and the fact that while some companies (Bumble Bee tuna) purport to be doing this, the feature did not work when I gave it a try for my map. Perhaps I could investigate this technology further in my next project and weigh the outcomes.
All that said, here is where my cynical side comes in – why would companies endeavor on this path other than the attempts to be “good people” and transparent entities? Of course we hope that’s enough, but where do these things fit into an economy based on such large scale…if at all. Jackson touched on this a bit in “Rethinking Repair,” but I’m curious to discuss more on how making these systems transparent can also make them sustainable, in both an ecological and trust-worthy sense moving forward.
This past week I attended the “Creating interactive, visual, data driven websites in WordPress” workshop hosted by theGCDI Digital Fellows. The workshop focused mostly on how to embed various forms of content on a WordPress page, and in turn, also helped to give me a better idea of the types of content I can create on the Commons for course assignments and research. One of the most important takeaways I had from the workshop was that embedding content is key for online publications – it keeps readers focused on the content at hand (stopping them from clicking out of your writing and getting lost in others tabs on the internet), and allows you to make more dynamic sites to showcase your research.
A few key terms outlined at the beginning of the workshop:
Embed: How to place content on your website directly to keep readers engaged and not link them elsewhere
Shortcode: An advanced shortcut that allows you to add features to your website that would normally require coding
Although the beginning of the workshop was tailored to individuals who had no previous background with the Commons and WordPress in general, they discussed types of privacy settings, general settings and page templates I found helpful:
In general, you can keep some pages in draft form and/or with certain privacy restrictions to keep your unfinished (or finished work) outside the public eye.
When you need to create group projects involving multiple users, simply go to your Dashboard -> “Users” -> “Add New Users” – users will then receive an invitation via email
It’s always good to consider the legal aspects involving some final WP publications, i.e. consider the ethical questions when posting social media research – you can often ask a user directly for their consent!
There are three types of templates you can choose from when creating a new Page on the Commons: Default (good for blogging), Teaching (good for setting up a course), Academic Portfolio (resume building)
We then moved into discussing WordPress Plug-ins and embedding for the majority of the workshop:
The automatic WordPress Block Editor set for Commons pages comes with a variety of pre-set plug-ins for embedding content like YouTube/SoundCloud/Vimeo/etc. You can find the list here!
Some sites/tech will not work this way, for instance Tableau. To directly embed these types of displays, try finding a shortcode on Google – or checking out this useful guide to Shortcodes on WordPress!
An example in action:
(Using Spotify embed – you can give a playlist for people to listen to while reading your research!)
Map of my refrigerator (and pantry)Zoomed into NY-area
Project Introduction
For my mapping project, I decided to “map my refrigerator(and pantry).” As someone with a big interest in food, I was curious to see where the everyday items I was consuming actually came from, and if my pantry/fridge is really as “global” as I think living in a city like New York. I had big dreams of mapping the journey from farm to distribution center to delivery route to grocery store (to delivery) to home, but after meeting with a Digital Fellow, realized my dreams may be too ambitious for my current mapping skills. We decided it could be a good start to begin trying to map origin points of an item, and that QGIS would be a good software from which to explore the visual aspects of such a simple data set…but I soon realized it would be much harder to find the data than i imagined..
Note: Since the project was completed on my computer on QGIS, I don’t believe it’s possible to post a link. Please let me know if it is!
Capturing Data
I began with the concept of mapping my most recent Fresh Direct delivery. However, after looking at a few items and doing some google investigation, I realized that a majority of the items were a dead end. I decided to expand my reach into a random selection of items in my fridge/pantry from differing sources (FreshDirect, H Mart, bought locally, etc.) and see where it would take me. To get a large enough sample set, I went with 20 items, and another list of 10 items that had interesting dead ends:
Initial data set (with dead ends)
I decided to categorize items by country and item type to see if there were trends between the two. However, I would some come to realize that a majority of my items were from the NY-area (which didn’t do much in terms of the country discussion). Finding the actual coordinates of the place of production/growing was quite difficult. For many products, it was impossible to find the locations where the items were made, but very easy to find their corporate office (i.e. Fuji Apples or Ocean Mist Farms spinach).
Surprisingly, the items that were advertised as straight from the source actually had no listed source at all., e.g. FreshDirect noted their 2-year aged Parmesan came from a “small cheese-maker in the Apennine hills of Emilia-Romagna,” but it was not possible to find a name/location of origin. Other companies, like Bumble Bee for tuna, even went so far as to have an individual can tracker, again to no avail – the can was listed as “Made in Thailand,” but was the fish? Not sure. A not so surprising find was that Trader Joe’s was not as transparent as they say, with almost no information on their packaging as to where the products were coming from.
For the items I DID find an origin location for (defined here as where the item was likely grown/canned/produced), I collected the relative long/lats for the map. What’s important to note here is the serious risks taken in the delineation of “origin” – companies could be lying, Google Maps could be wrong (especially in the case with the Tahini site in occupied Palestine), more complex products (such as take-out) are really the result of dozens of products, and sometimes I had to take a guess that the factory I received the product from was the one closest to my delivery point. I tried to illustrate these complexities in the “Notes” categories listed above.
My final points for mapping were:
Cleaned data for QGIS map
Mapping with QGIS
After a quick introductory lesson from a CUNY Digital Fellow (shouts out Rilquer!), I went into the mapping process. I spent a solid hour trying to figure out what was wrong with my coordinates when half of them showed up in Antarctica, and quickly realized I needed to clean up my longitudes/latitudes. Then, in combining them with a map of states/countries downloaded from an open source site, Natural Earth Data, they magically appeared! To make it more interesting for people to read, I decided to label them with the item name, and coordinate the colored points by category type – which as denoted earlier, didn’t do much in terms of analysis. I also played around with the “Heatmapping” feature to try and delineate areas trends in purchase locations. Again, given there wasn’t too much variety in my sample set (yes I am a lazy grocery store delivery person), it was not surprising that Fresh Direct was mostly local to NY.
What did strike me in seeing the items visualized was just how many of my products come from the New York area – and how one of my furthest sourced items was meat (yikes). I did some research to see if that was the case, or if that was a matter of FreshDirect actually being fairly “fresh” and “direct.” What I found was that NY is actually quite a hot spot for growing, particularly around vegetable and yogurt – which was evident in my map! That said, it’s also important to think about how grocery stores may reduce their products to local/closer regions in order to save money.
Source: Farm Bureau New York
Where to go from here
I would love to do more with this concept for my final project. First off, I could have worked with a much larger data set in order to really pick apart some trends, but was restricted in terms of what was actually in my house and how easy their production locations were to find, alongside the general timeline/scope of this introductory project. Moving forward, I’d love to consider mapping certain aisles in a store (perhaps those labeled “ethnic), a whole section of my pantry, a certain recipe, a larger home’s refrigerator, etc. The possibilities are endless!
I also think an ultimate goal is trying to map the routes that these products are taking and start to get a sense for how “local” local products really are, or better visualize where my food is coming from. In doing so, I would need to learn a bit more about how to map routes, and whether QGIS is the right software for the task. And more questions – would I also want to visualize the date behind the route’s environmental impact, cost, etc.? The questions are endless!
Using the project Torn Apart / Separados to (re)define Digital Humanities
The Digital Humanities is a field that centers on creating and disseminating collaborative, open sourced and interdisciplinary research projects that often adhere to a particular political agenda. Projects are digital in nature and almost always include pedagogical instructions and a description of methodology. In terms of the political effectiveness, DH believes that “small acts of recuperation” can create “building blocks to larger collective action.”
Torn Apart / Separados situates the Digital Humanities as a discipline using digital tools to (re)produce/(re)position the data informing and derived from current events/historical sources to inspire change and action. Here, the topic is ICE’s financial regime in the US, and an implicit call to action through “the data and visualization intervention… of culpability behind the humanitarian crisis of 2018.”
Digital Humanities is a discipline focused on exploring humanistic “data” -in this case both qualitative and quantitative- and creatively interprets and visualizes data from a broad range of sources and media types (institutional archives, public reporting, and social media posts are all examples of “data” for DH). DH is focused on translating a data set into a living “product” or visual/textual project that grows and updates alongside the evolution of the study and reporting on the topic in question. (Here, the project has two volumes, and both iterations remain visible online with details on the changes per volume.)
It’s also highly interdisciplinary — bringing together a range of professionals and types of reporting to enhance the spectrum of research methods utilized and types of data explored. These participants, practices and sources are also highly documented in a way that promotes further investigation into the topic at hand and the technology platforms themselves. In turn, the discipline inspires researchers and everyday people to consistently question traditional modes of information sharing and lean towards a methodology that is more relational and expands context across space and time.
DH projects inherently call into question the systems used to gather information for inquiry through their (re)positioning and (re)purposing of data. Here, they explicitly ask us to question the “truth” behind reporting and numbers released by governing bodies and major companies in regards to ICE funding and operations. As a result, a unique characteristic of the discipline is that it is inherently contradictory as it relies on the systems of information it seeks to dismantle and reposition (ICE records).
While “open access,” (aka available free/online) the discipline also seems restrictive to cultures and individuals who value numerical reporting and textual analysis over more subjective forms of knowledge and testimonial like oral history. However, this project is particularly visual and does appeal to individuals who might be more literate in data vis. Perhaps this is the shift from “humanities” to “digital” humanities? For instance, here, we are asked to examine and explore financial reporting, rather than the stories behind the individuals deported.
To what end? That’s up for debate, as DH projects call for others to step in and expand on the project as a way of keeping the “story” alive. In this project, there is no specific call to action other than a new addition to Volume 2 that highlights “Allies” in the fight to expose ICE. Transparency is clearly a major player in Digital Humanities, with the call to action being making information more widely accessible in a way that advances institutional critique, rather than promoting some specific IRL action towards change-making. That said, I look forward to discussing the ways in which this might succeed/fail with the class!
Need help with the Commons?
Email us at [email protected] so we can respond to your questions and requests. Please email from your CUNY email address if possible. Or visit our help site for more information:
