My #HuntLibrary: Using Instagram to Crowdsource the Story of a New Library

[Updated to reflect open-source availability. May 13th, 2013]
 
Introduction

North Carolina State University opened the James B. Hunt Jr. Library in January of 2013, creating a heart for our Centennial Campus that defines the research library of the future. My #HuntLibrary was created as a platform to foster student and community engagement with the new building via social media imagery and to preserve and archive these images as part of the record of the Hunt Library launch. My #HuntLibrary is a Ruby on Rails application that harvests images from Instagram and provides several browsing views, mechanisms for sharing, tools for users to select their favorite images, an administrative interface for moderating images, and a system for harvesting images for inclusion in the NCSU Libraries digital archives. Built according to the principles of “responsive design,” My #HuntLibrary is usable on mobile devices, tablets, desktops, e-boards, and the massive MicroTiles displays in the Hunt Library.

In the three months since the launch of My #HuntLibrary (coinciding with the opening of the Hunt Library building), we received nearly 1700 images from over 600 different users, over 6800 “like” votes, and over 53,000 “battle” votes. This post will detail some of the risks involved with the project, the technical challenges we faced, how student engagement strengthened the project, and the potential benefits of giving students and community members a voice in the documentation of the Hunt Library.

The code that drives My #HuntLibrary has been extracted into an open-source Rails Engine called “lentil” that is available on GitHub.

My #HuntLibrary front page

My #HuntLibrary

Planning for Risk

Most projects carry some level of risk and My #HuntLibrary was no different. It was difficult to predict the level of engagement we would be able to achieve with various application features. The timeline for development was short, carried a firm deadline (due to the need to launch alongside the new library), and included work with several technologies that were new to the development team. Additionally, the application relied on a third-party API that could change at any time. In order to mitigate project risks, we structured the project around goals with short and long (and more speculative) timelines that would each individually justify the project effort.

  1. Utilize social media to increase engagement with a new library

Social media engagement with students was a linchpin of our opening strategy. Before the Hunt Library came online, NC State students already had a high degree of ownership over existing Libraries spaces and we sought to extend that to our new library. My #HuntLibrary could contribute to that sense of ownership by providing a platform for users of the new library to document and share their experience, learn about the experiences of their peers, and to collectively curate the images using voting tools. Furthermore, My #HuntLibrary is an opportunity for staff to learn about important and unexpected uses of the building during the critical post-launch period.

  1. Provide a mechanism for students to contribute to digital collections

We felt that the Hunt Library opening could be an opportunity for students to add their voices to the documentation of campus history that is preserved in our extensive digital collections. My #HuntLibrary could also allow us to leverage social technologies to diversify the perspectives reflected in our archival collections. This is our first major social media preservation effort and we hope that our project, along with others (such as GWU Libraries’ Social Feed Manager, or the State of North Carolina’s Social Media Archive), may begin to contribute possible answers to several questions related to social media archives, including:

  • Can we utilize existing streams of social media content to integrate additional student perspectives in our documentation of the history of their university? Does this enhance our special collections?

  • Can an invitation to participate in core library practices, such as the development of special collections, serve as an effective engagement strategy?

  • What is the research value of social media collections? How does this value vary based on media, users, and harvesting methods?
  1. Explore new technologies

The developers involved with the project created a support structure that included pair programming, code reviews, and tutorial sessions that mitigated many of the technical risks, including the integration of new software frameworks and libraries and the coordination of work on a tight schedule. This project also provided an opportunity to learn more about the design of interfaces for the large-scale displays described later in this article.

Student Engagement

Although we knew it would be possible to utilize the Instagram API to collect and display photographs about the Hunt Library, we needed to have a reasonable expectation that people (and students in particular) would participate in the project. This question hinged on the likelihood that a person would tag a photograph of the new library with a hashtag that would allow us to capture it. The Libraries had previous experience trying to engage students through Twitter around the question “What are you doing in the library right now?” We looked back on that project’s limitations to inform our engagement strategy. The chosen hashtag (#whyncsulib) was unique, but in order to answer our question, students had to be aware of the hashtag and willing to deviate somewhat from their normal social media communication patterns. However, we found that it was already common for students to use the tag #DHHill to visually depict their activities in our D. H. Hill Library on Instagram. 

Example #DHHill Instagram images

Example #DHHill Instagram images

Assuming that students would continue this tagging behavior at the new library, we chose the hashtag “#HuntLibrary” in hopes that it would see widespread adoption regardless of the level of awareness of our project.

As we began to design the application and develop a social media plan, another milestone in the project came with the opportunity to present the idea to actual students. The NCSU Libraries Student Advisory Board is charged with providing guidance and input on the programs and services the Libraries offers. This regular open meeting (fueled by free food) allowed us to collect feedback on specific questions about the project (e.g. do students want images to “Battle?”). The feedback from this presentation varied widely, from useful (e.g. roughly two-thirds of the students present had Instagram installed on their phones and yes, they want to battle) to unsanctionable (“If you want cute photographs you should let us bring cats into the library”). However, the general reaction from the students was that it seemed like a good idea, and we continued work with increased confidence.

The Student Advisory Board meeting also led to another breakthrough: our Director’s commitment of funds to award an iPad Mini to the photographer of the best image. Prior to the Advisory Board meeting, our only participation incentive was an assurance that the best photographs would be ingested into the University’s permanent digital archives. While this is a thrilling idea to a roomful of librarians, we were uncertain that students would have the same reaction. Perhaps unsurprisingly, when our Director asked the gathered students if they would take more pictures if there were an iPad Mini at stake, the students were unanimous in their response. Although we later learned in usability tests that students reacted very positively to the idea of contributing to their University’s story, the tablet prize gave the project a focal point, and the contest became the cornerstone of our student engagement strategy.

Display Technology

The NCSU Libraries’ vision is to be NC State’s competitive advantage. This vision is often operationalized by putting cutting-edge technology in the hands of our students and faculty. For the Hunt Library, we made a strategic investment in large-scale, architecturally integrated visualization spaces such as ultra-high definition video walls and virtual environment studios. These visualization spaces serve as large canvases to reflect the research and activities of our campus in new interactive ways. The Hunt Library is, in short, a storytelling building.

We anticipated that My #HuntLibrary would produce a visually compelling record of the new library, and so we chose to display the photographic activity in one of the library’s most accessible visualization spaces: the iPearl Immersion Theater. The Theater features a curved video wall that is twenty-one feet wide and seven feet tall. The wall uses Christie MicroTiles, a modular display system based on LED and DLP technologies that gives the wall an effective resolution of 6824 pixels by 2240 pixels. MicroTiles feature high color saturation and a wide color spectrum, making them ideal for Instagram photographs of the colorful library. A key part of the technology behind the MicroTiles is a Christie Vista Spyder. The Spyder is a hardware-based video processor that allows for 12-bit scaling. This upsampling capability was important for our application, as it allowed small (612 pixels square) images to be enlarged to two-foot images in the Theater with very few noticeable compression artifacts. 

Viewing My #HuntLibrary in the Immersion Theater. Photo by Instagram user crmelvin14.

Viewing My #HuntLibrary in the Immersion Theater. Photo by Instagram user crmelvin14. 

As a public, physical space, the iPearl Immersion Theater allowed us to create embodied and shared user experiences that were fundamentally different from the web and mobile views of My #HuntLibrary. The Theater is a semi-open space near the entrance to the library, adjacent to an expansive reading lounge. The video wall installation had an attractive presence that invited passers-by inside to examine the images. Once inside the Theater, the content could be appreciated more fully by moving around in the space. Standing close to the wall enabled the user to see more detail about a particular photograph while moving farther away gave an impressionistic sense of the library’s spaces. While dwell times for the installation were sometimes low because users often dropped in for a moment before heading to their intended destination, seating in the Theater allowed for a more leisurely viewing experience as new photographs rotated into the display. Small groups of people gathered in the Theater to discuss the merits of their favorite photographs, point out their own photographs to their friends, and engage in conversations with strangers about the images.

Responsive Web Design

With the large MicroTiles displays in the Hunt Library we now face the challenge of designing for very small (mobile device) and very large displays and many sizes in between (tablets, laptops, desktops, e-boards). The growing popularity of responsive web design techniques have helped developers and designers meet the challenge of building applications that work well on a wide range of device screen sizes. Responsive web design generally means using a combination of CSS3 media queries, fluid grids, and flexible images to progressively enhance a single web design for optimal display and use on a wide range of screen sizes and devices (Marcotte 2010). Most of the discussion of responsive design centers around building for devices ranging from phone-sized to desktop-sized displays. However, there is no technical reason why responsive design cannot work for even wider ranges of display sizes.

Our final design for My #HuntLibrary includes two different responsive designs, one of which supports mouse and touch interactions and display sizes ranging from phones to desktops, and another for non-interactive public display of the photographs on displays ranging from large eboards to more than twenty-foot wide Christie MicroTiles arrays. Our decision to build two different responsive designs for the smaller and larger sets of displays has more to do with the context in which these displays are used (personal, interactive devices versus public, non-interactive displays) than any technical limitations imposed by responsive web design techniques. In our case, the design of My #HuntLibrary for phones, tablets, and laptop and desktop computers has features to support interactive browsing, sharing photos, and a photo competition “Battle View” for people to compare sets of images and pick their favorites. These features would not translate well to the Libraries’ larger public displays, which range in size from a large eboard to huge Christie MicroTiles video walls, and which are, for now, mostly non-interactive. It made sense to develop a different view optimized to support a non-interactive display of the My #HuntLibrary photos. For the eboard-sized and larger public displays we developed a grid of images that are periodically replaced by new images, a few at a time.

Mobile view of My #HuntLibrary.

Mobile view of My #HuntLibrary.

My #HuntLibrary on Christie MicroTiles in the Immersion Theater.

My #HuntLibrary on Christie MicroTiles in the Immersion Theater.

Collecting Social Media

Although the initial development push was heavily focused on the core data management and display infrastructure, the longer-term goal of content preservation (for the sake of historical documentation rather than personal archives) influenced most aspects of the project. In particular, we have attempted and are continuing to address three major preservation-related themes: harvesting, crowdsourced curation, and legal clearance.

For short-term use of the images, we harvest only the metadata, leaving the images on the Instagram servers. Clearly, for long-term preservation we would need to collect the images themselves. This harvesting is complicated by the necessity to declare an arbitrary “break” from the source materials, at which point any changes to the metadata (or removal of the images) would not be reflected by our system. We are currently developing a milestone-based harvesting schedule that takes into account both the length of time the image is in the system and the submission of a donor agreement.

While we are currently planning on collecting all “#huntlibrary” images, we are very interested in the potential to allow our users to influence the selection process for certain parts of our archival collection. In order to test and support this goal, we developed two voting tools: individual image “like” voting and this-or-that “battle” voting. Our hope (which early usage metrics seem to support) is that we could use the data from these tools to select images for preservation, or at least to promote a subset of preserved images, that reflect the interests of our community. In addition to improving our selection processes, this may be an opportunity to promote archival selection as a student engagement tool by promoting opportunities for students to influence the historical record of their own experiences.

Image battle interface.

Image battle interface.

Finally, we worked with a lawyer and copyright specialist at our library to develop a donor agreement that was short and clear enough to be submitted as a comment on an image. Instagram users retain rights to their own images and thus the ability to grant the limited rights that we are requesting. Furthermore, the use of the Instagram comment system will allow us to automate this process, provided that we are responsive to takedown requests.

Conclusion

In the three months since the launch of My #HuntLibrary (coinciding with the opening of the Hunt Library building), we received nearly 1700 images from over 600 different users, over 6800 “like” votes, and over 53,000 “battle” votes. In addition to these measures of user contributions (of either images or vote-based reviews), My #HuntLibrary recorded 135,908 pageviews from 10,421 unique visitors (according to Google Analytics) during this period. Furthermore, the project was regularly cited by students, staff, and institutional partners on social media channels, and was featured (with an emphasis on historical documentation) during the Hunt Library Dedication events.

The evaluation of the archival components of this application will take place on a longer timeline. We are currently extending the long-term content harvesting features in order to support these activities in a more automated way. We have received several indications of the value of pursuing image preservation features, including surprisingly enthusiastic reactions to questions about the preservation of images from students taking part in a My #HuntLibrary user study. As a particularly encouraging example, when an undergraduate student contributor to My #HuntLibrary was asked “How would you feel if one of your Instagram photos were selected by the library to be kept as a permanent record of what students did in 2013?” they responded, “I would be so excited. For me, I think it would be better than winning an iPad.”

About our guest authors:

Jason Casden is the Lead Librarian for the Digital Services Development group at the North Carolina State University Libraries, where he helps to develop and implement scalable digital library applications. He is the project manager and a software developer for “My #HuntLibrary,” and has served as a project or technical lead for projects including the Suma physical space and service usage assessment toolkit, the WolfWalk geo-enhanced mobile historical guide, and Library Course Tools.

Mike Nutt is a Fellow at NCSU Libraries, where he leads a strategic initiative called “Networked Library: Marketing the 21st Century Library.” He is the product lead for My #HuntLibrary, and also facilitates content strategies for the large video walls in NC State’s new Hunt Library. He founded the University of North Carolina at Chapel Hill student group Carolina Digital Story Lab and was a research assistant at the UNC-CH Carolina Digital Library and Archives.

Cory Lown is Digital Technologies Development Librarian at North Carolina State University Libraries where he works collaboratively to design and develop applications to improve end-user resource discovery and use of library services. He has contributed as a developer and/or interface designer to a number of projects, including My #HuntLibrary, WolfWalk, QuickSearch, and the latest version of the library’s mobile website.

Bret Davidson is currently a Digital Technologies Development Librarian at the North Carolina State University Libraries. Previously, Bret worked as an NCSU Libraries Fellow on visualization tools and resources to support the new James B. Hunt, Jr. Library. Prior to becoming a librarian, Bret was a music educator in the public schools of Pennsylvania and Illinois, as well as a performing musician with the River City Brass Band in Pittsburgh, PA.