Final Reflection: Middlesex County Oral History Project



This semester I used the content management service Omeka to build an oral history archive of Middlesex County, Virginia. The idea for the project began in 2014 when I was home for Christmas. My mother saved an article from our county’s newspaper about a group of University of Florida undergraduates who did field recordings of Middlesex residents. After speaking with a local museum employee, I learned that there were no plans to make the recordings available online. My goal for the Middlesex County Oral History Project was to make these recordings accessible, and to encourage residents to engage with their own history by directly submitting stories, photographs, and comments of their own. Hopefully, as the collection builds, a larger picture of our community will emerge from these micro stories that will prove valuable for residents.

After speaking with the director of the University of Florida program as well as Middlesex Museum staff, I created a project proposal for the scope of LBSC 708D. Below is a comparison of promised deliverables vs. what actually happened by the project’s deadline:


After learning that the UF field recordings would not be available until June, the immediate need became creating content and access. I conducted an additional oral history using the iphone app “Call Recorder” and created Youtube videos (at Trevor’s suggestion) to engage site visitors in interviews that were in some cases over an hour long.


Friends and family involved in user testing had many helpful suggestions, such as changing the homepage layout, the order of item pages, reporting broken links, reporting item submission bugs, and in general telling me if something “looked weird”.

Although I was happy with the site’s development, my worry was that Middlesex residents wouldn’t submit oral histories either because it was too time consuming or seemed too intimidating. I also knew that I needed more content to encourage people to submit, but the very act of interviewing people gave the oral histories a professionalism that seemed unapproachable. Fortunately I had the pleasure of hearing Todd Wemmer, a Communications professor at Endicott College, discuss his project Photos Die at the 2015 Personal Digital Archiving Conference. wemmer

After setting up a voicemail service through Skype, Wemmer encouraged users to call and leave a short oral history that he would later capture and share through Soundcloud. So far his strategy has been successful, especially with elderly storytellers who are not familiar with digital environments. By adopting Wemmer’s idea as an additional method of storytelling, my hope is to appeal to a wider group of Middlesex County residents on their own terms.

Although I’ve completed this project for the scope of our class, the long-term success of the project will be based on the number of site visits and the amount of content submitted. This summer, I plan to begin a promotional campaign for the site, including a write-up in local newspaper The Southside Sentinel and public presentations at both county libraries. I will also promote the site to teachers at Middlesex High School who might be interested in teaching a unit on oral histories.


Although I’m enthusiastic about the potential of this online archive and have spent many hours on its development, the website’s appearance seems simple and sparse. I’m eager to start promoting the project, collecting content, and improving the archive’s value for community members.

All About That Bot


According to Wikipedia, bots are software applications that run automated tasks over the internet. Although bots can be written for all types of things on the web, many of the ones I encountered before this week’s readings were spam advertising. (If you were using a desktop in the third quarter of 2014, 41% of the online ads you were seeing were bot-generated.)
The number of bots is also growing rapidly on Twitter, whose open API makes them especially easy to script. The social media company released a statement in 2014 saying that 23 million users were automated accounts, or bots. Don’t assume that they are all trying to get your money, though. In this post, I’ll explore some of the good guys.
The Protest Bot
In Mark Sample’s article, we learn that bots can be written as a form of protest. His bot @NRA_Tally pulls real victim numbers, locations, and types of weapons from mass shootings that have taken place in the United States in the past thirty years, and matches them with one of 14 stock NRA responses. It’s also participatory, so users can submit generalized versions of documented responses and Sample will add them to the bot code. Although Sample outlines several characteristics a protest bot must follow, the most interesting one by far is the aggregate power of the bot:
It is the nature of bots to do the same thing over and over again, with only slight variation. Any single iteration may be interesting, but it is in the aggregate that a protest bot’s tweets attain power. The repetition builds on itself, the bot relentlessly riffing on its theme, unyielding and overwhelming, a pile-up of wreckage on our screens.

So in the case of @NRA_Tally, the point is to literally spam you with these examples so that they’re unignorable. As he says, the purpose of the protest bot is to “create messy moments that destabilize narratives, perspectives, and events.”
The Cultural Institution Bot
In our line of work, bots have been created to promote cultural institutions by displaying their wares. MOMA created @MuseumBot to tweet images of their collections two to three times a day. These tweets are relatively simple, consisting of an image, sometimes a title, and a link to the item within the online collection.
Lubar advocates for bots to fill gaps in access and expose patrons to the agency behind what items make it into an exhibition. Some of his suggestions include tweeting things in storage that aren’t on display, or images of items that were deaccessioned. I find this idea incredibly liberating because it takes the stark polarity of “will this be seen or not be seen?” out of appraisal decisions, and helps shine light on a growing percentage of dark archives.
The Poser Bot
I decided to ask a bot for their opinion on the matter. @OliviaTaters is a bot that was made accidentally by former Colbert Report writer Rob Dubbin. Basically, Dubbin was trying to exploit new language trends in teenage tweets, words such as “actually,” “like,” emoticons, etc. Although the bot was meant to put pressure on these expressions, @oliviatater’s tweets were often funny, witty, or touching, and she is followed and speaks with many teenagers today. To hear more, listen to this hilarious TLDR podcast.
Here’s my tweet to Olivia:

As you can see, she wasn’t very helpful, although I guess you could tease out meaning about historically-set videogames outpacing cultural institutions and denial being a part of that relationship? I admit, it’s a bit reaching, but perhaps the ambiguity of language is what makes @oliviataters so compelling. Tim Sherrat makes the point that the bot’s failings are themselves valuable, because they show us what it means to be human and encourages us to analyze and critique the medium.
The Artistic Bot
Rob Dubbins describes #oliviatater’s tweets as a kind of literary art in the interview I cited above. @TwoHeadlines , a Darius Kazemi creation that pulls top headlines from Google News and splices them together, is another example of a bot that was originally created for one purpose, but suddenly started producing literary art.
One tweeter viewed such headlines as a “magical realist Late Capitalist dystopia.” Kazemi agrees that they could be described as near-future-late-capitalist dystopian microfiction, sort of in the vein of Infinite Jest, where nations, sports teams, and celebrities become the same thing. These headlines reveal the dangers of capitalism and privatization when these trends start to affect people and public entitites.

There’s also bots that make visual art. @Pixelsorter uses an algorithm to resort the pixels of a picture according to hue, brightness, etc. After one failed attempt, I successfully got the bot to glitch my dad’s picture:


The prolific Kazemi also made @reverseocr , which draws lines until optical character recognition software interprets a word out of them.


These examples have made me re-think the bot as not just an annoying spam machine. It has the ability to protest, advocate, soothe, or to inspire. It may even push other sites to make their API available (I’m looking at you, TripAdvisor!), although that may be too optimistic.

Does anyone else have a favorite bot that would fall into one of the categories, or a humanities related one they’d like to share?

On Context


In this post, I will not be addressing what is and isn’t an archive (Joe’s throwing down on that), but for full disclosure, I will be using the terms “archives” and “digital collections” interchangeably when referring to a grouping of materials.
What I will do in this post is compare features of digital archives highlighted in the readings, and discuss how they are affecting the context of digital collections in new and awesome (?) ways.

The Hyperlink
McGann is right to say that hypertext is redefining how users interact with archives in the digital environment. Instead of needing books to analyze books, digital material (born-digital and surrogates) allows the “white sea of paper” to recede. Researchers are no longer limited by geography or representation when deciding what objects they’d ideally put in dialogue with each other . Hypertext has the ability to aggregate dynamic objects together to inform or comprise a collection, and although Theimer implies that this violates the context of these collections, McCann argues that hypertext allows us to create richer context for our collections that was never before possible through this digital tool. Of course, hypertext is so easy that it allows anyone to be able to create a collection, not just the “professionals” that respect archival theory, have an HVAC compliant preservation environment, or the budget or reputation to acquire material, and so this loss of control may be seen as threatening (a familiar theme for this class, I know).

Visualization tools

In his article “Disrespect de Fonds,” Jefferson Bailey introduces us to the Series Browser Visualization tool from the Visual Archives Project. With this tool, the context and codex-like purposes of a finding aid are replaced by networks and inter-linkages that communicate the relationships and size of hundreds of record groups, all visible to the user on a single screen. As Phillips points out in her article, this capability creates some questions. If we can make this scope accessible for researchers, should we just take everything? Will changing what users can get affect what they actually want? Although Phillips is specifically discussing appraisal, we can see it as a context issue as well. Selection of what to keep is the act of an archivist indirectly creating a context for the user; the responsibility of deciding if something is trash or treasure is, in my opinion, one of the most stressful aspects of the job. Through visualization tools, appraisal could potentially become a thing of the past; a hindrance to context that was a product of its time, much like Bailey argues fonds was a product of its time that has also become obsolete in the digital age.

Deep Freeze

In Schmidt & Ardem’s look at the Susan Sontag Archive, they discuss UCLA’s decision to allow researchers to check out Sontag’s laptop as an access model. Archivists treated the laptop with a deep freeze, a process that essentially preserved the laptop at the bit level so that it can be explored by a user but not changed. The authors’ question whether preserving context in this way effectively tells the whole story:

When we search, we may feel as if we can finally, totally, and quickly find everything we want to know about a particular aspect of Sontag’s work or life, but the very ease of that process makes our misguidedness plain. What we are “finding” is, of course, nothing more or less than particular words in a particular order.

Having the writer’s laptop in your hands makes you think that you have it all, but as we know, even something as ubiquitous now as text searching leaves out semantics that a computer command just can’t register at our current technological capability.

Final Thoughts

Even though there are still many challenges, looking at these features makes me incredibly excited about the potential for digital collections to tell richer stories when freed from aspects of appraisal and geographical limits, for finding aids to be reimagined as visual maps vs. textual ones, and for true original order to be captured at the bit level. But if i’m honest, it’s a little scary too. What if we’re eschewing things we shouldn’t be? Bailey’s article consoled me that there was never a “true” way to process things, just a way convenient for its time. But does anyone else have a nagging, apocalyptic side that sees the bubble bursting after we burned the book that got us here? Maybe this Sagan quote will warm you.


Middlesex County Oral History Project (Digital Proposal)

As a part of Virginia’s Middle Peninsula, Middlesex County is bordered to the north and south by the Rappahannock and Piankatank Rivers, and to the east by the Chesapeake Bay. The earliest record of the County is from explorer John Smith’s journal in 1608, when he and his crew ran aground near the mouth of the Rappahannock River by the town currently known as Deltaville. Smith was excited by the large amount of fish in the area and began spearing them, only to be stung in the arm by a cow-nosed stingray. The injury was so painful that Smith’s men prepared a grave for him, but a doctor in his crew successfully treated the wound and Smith survived. “We called the Isle Stingray Isle after the name of the fish,” the explorer wrote, and the name survives 400 years later as “Stingray Point”.

Middlesex’s geography continued to play an important role in the County’s development and culture, creating opportunities in the crabbing, oystering, and ship-building industries, while safeguarding the county from the destruction of the Revolutionary and Civil Wars (our courthouse still holds it’s original records from 1679). It’s geographic isolation, and the effects of several large hurricanes that destroyed it’s shoreline and farmland in the 1930s, slowed the County’s development over the 20th century, while preserving a distinct regional identity.
Although the county is rich with stories from the Tidewater, efforts to collect oral histories have been done sporadically in print form through local newspaper The Southside Sentinel, and several books available through the Middlesex County Historical Society’s website.
The Middlesex County Oral History Project (MCOHP) is an effort to create an online archive of user-contributed oral histories in audio, video, and text formats.


Following in the footsteps of regionally focused oral history archives such as Philaplace in Philadelphia, the Maria Rogers Oral History Program in Boulder, and Old Dominion University’s Tidewater Voices project , MCOHP will serve as a resource for historians, linguists, educators, and most importantly, Middlesex County residents.

The site will be hosted by Omeka, an established content management system for many small cultural institutions. Residents will be able to upload oral histories, transcripts, and photographs through the website, and Museum staff will process the digital files with Dublin Core description and local tags such as names, places, and subjects mentioned in the history. Digital files and metadata will be preserved and stored using Omeka’s cloud-based server, and the Museum will keep additional copies of all material on their local server.
The Museum will work with Middlesex High School teachers to create lesson plans for the site that meet Virginia’s Department of Education Standards of Learning, and will provide user-friendly documentation for recording oral histories, such as a sample permissions agreements, interview questions, and suggested recording formats. Several workshops have been planned with the Deltaville and Urbanna Public Libraries to promote the site and educate the public on responsible oral-history creation.

As submitted histories increase, the MCOHP may organize them into topic or town-based collections. Select histories will be used in conjunction with Museum exhibits, mobile tours, and as content for the Museum’s blog. In the future, the Project would like to move it’s website and collections to a locally based server for increased administrative  and creative control.

The Middlesex County Museum and Historical Society hope that this archive will promote Middlesex County as a research destination. We also hope it will become a source of pride for our community members as it facilitates a deeper understanding of our home, and creates connections amongst residents as they explore the variety of voices that make-up our County’s identity.

What TripAdvisor Says About America


Bizarre, contentious, and extremely popular, Mount Rushmore National Memorial has been etched into the U.S. identity since it’s construction in 1877. According to the official site, it “[symbolizes] the ideals of freedom and democracy.” The National Park Service’s page for the memorial features this quote from Rushmore’s architect Gutzon Borglum:

The purpose of the memorial is to communicate the founding, expansion, preservation, and unification of the United States with colossal statues of Washington, Jefferson, Lincoln, and Theodore Roosevelt.

“Colossal” couldn’t be more accurate. With 20-foot high Presidential faces carved with dynamite, this American landmark in the desert of South Dakota is a landmark so ambitious that even in the desert it sticks out. But Mount Rushmore’s history is as equally difficult to ignore , raising uncomfortable questions about it’s preeminent place in our culture. Borglum, himself, served as a member of the Klu Klux Klan , and the Rushmore property (which at one time had been willed to the Sioux Nation) was taken back by the government in 1874 when gold was discovered there. In 1980, the U.S. Supreme Court ruled against the federal government in favor of the Lakota tribe, but the case remains unresolved.

This paper will analyze user’s reviews of Mount Rushmore on the social media site Tripadvisor to explore perceptions of the memorial as it relates to the United State’s national identity. With ten years of data, I will use visualization tools such as Voyant to analyze linguistic trends that rise to the surface, whether these trends change over time, and if certain words correlate to negative or positive reviews (only 219 of the 3,508 reviews were given an “average” or below rating). Some of the questions I hope to answer include: (1) Are visitor’s aware of the monument’s history, and if so, does it influence their rating? (2) What language correlates to negative and positive reviews? (3) Do linguistic trends change over the decade? (4)After analyzing these subsets, what can be concluded about how reviewers interpret their own national identity?

Those that came before
Using social media to analyze how the public defines and creates meaning has been an emerging field of scholarship that this study hopes to continue. In “My Tripadvisor: Mining Social Media for Visitor’s Perceptions of Museums vs. Attractions,” Elizabeth Mauer used Tripadvisor comments to analyze the expectations and perceptions of museum visitors in order to re-think how museums can better represent themselves. In “Trip Advisor Rates Einstein,” Trevor Owens shows that the creation of meaning through social media is recursive; a place to record reactions, while simultaneously providing a ‘frame’ that influences how new visitor’s interpret the statue. Bryan Routledge performs computational linguistics on social media to break open language trends and what they say about society. As an example, he and a team of researchers used Yelp reviews to show that expensive restaurants are most often described through metaphors of sex, while cheap restaurants were described through metaphors of drug abuse and addiction. These scholars have used an emerging corpus of data to make important statements about society and meaning-creation.


Methodology                                                                                           Unfortunately, Tripadvisor will not allow API access for academic research or data analysis, so I plan to sample 25-30% of the reviews for the purposes of this class. On the up side, Tripadvisor allows you to sort by date and reviewer ranking, so getting at the data from different angles will be relatively easy. Through the text-mining tool Ventura, you can search by specific words and also see overarching trends. I plan to use both methods on three different sample sets (negative reviews, positive reviews, and a span from 2004-2015), in order to answer my proposal questions.

It is my hope that this study benefits a cross-section of disciplines in the social sciences, contributes to emerging digital scholarship, and gives us insight into how we reconcile the sometimes contradictory narratives of a monument’s history and what it is supposed to represent.