Visualizations for the visuals

Curarium is essentially a huge amalgamation of some exquisite data — paintings, photographs and visual media content. While developing a platform for this extensive data-set, we noted that it’s purpose is not just what this data represents, but also the data itself. And how do we intend on presenting this data? Enter, visualizations!

A distinct advantage when dealing with huge data-sets is the population and representation of it, in varied colors and schemata – which make for elegant visualizations. But an efficient one requires to go beyond just the pretty. So the first challenge presented itself in the form of short-listing visualizations, that would fit into the frame-work of explorations, collections and spotlights within Curarium. They were to be navigable, embeddable and fulfil requirements of functionality. Keeping this in mind, we finalized a couple of them – thumbnail views, tree-maps, time-lines, geo-spatial maps and a few more. In the selection of each of the visualizations, we had to consider the use-cases and make them application specific. For instance,

  • In the thumbnail view, the different visual media are present in the form of thumb-nails. This was a more obvious choice in terms of actually being able to view the content that was being explored and the aesthetic appeal from a user perspective (humans seem to like to know what they’re getting into before-hand).



  • The tree-map plays out rather wonderfully through the range in color spectrum and increasing sizes of each individual record. An organized treat for the eyes is probably an appropriate description.
  • The integrated time-line with geo-spatial maps provide a functional representation of a record over space-time. During exploration, this would render a huge deal of context when it comes to the questions of when’s and where’s, and establishes a former-latter relationship among the records themselves.

Each of the picks were unanimously voted for trial, testing and application into the website. Presenting data that could easily be referred to as “my precious” requires tremendous care and a severe control in excitement. So in process of the make-over from a dump of JSON files to carefully picked beautiful visualizations, we are awaiting the perfect “Voila!” moment.

Curarium and the “Digital Humanities”

Students in the Humanities with so-called ‘digital’ interests are often attacked, not only by professors, experts in pre-digital research and publication methods, but more surprisingly, by fellow students. There is still a sense that Digital Humanities is simply a term for research done broadly by a computer rather than deeply by a human being. While this, certainly, was one of the first definitions of the term, it is by no means the only, and definitely no longer the most relevant.

Digital humanists are interested in refined and up-to-date modes of communication with culture.

Why write a beautifully-researched publication if no one will ever read it? 

And more to the point, How can we get the culture that we study interested in what we have to say about it?

Enter the role of the digital. We humanists can engage with society, in part, by beginning effectively to use the tools of communication that have become part and parcel of contemporary life. The AI substituting the organic is not and never has been an end-goal our work at metaLAB. The end-goal of our work could be a book on cultural studies that is created in collaboration with artists and designers and that rethinks the readers’ relationship with text. The end-goal of our work could be a chapter of a dissertation or book that literally reaches out to society, through a physical or digital installation or a cultural intervention that allows the community to enter into conversation with the work.

Or the end-goal could be something like Curarium.

Curarium is a platform that fills a digital hole in the web. It hopes to meet the goals of museum staff, researchers, teachers and enthusiasts who are looking for a user-friendly and intuitive place to import, search, view, organize, annotate, collect, and, perhaps most importantly, converse about digital representations of cultural artifacts.

The concerns of the digital humanist, then, is twofold, and seeks to bridge the two worlds without giving preference to either:

(1) What can we humanists do in the archives that we could never do online? Let’s not lose touch, or let our students lose touch, with the importance of the physical search, the in-depth study, the detail.

(2) What can we humanists do digitally that we could never do in a book? Let’s not forget that the internet offers new modes of looking, of story-telling, of broadcasting, and let’s not be suspicious of technology as a means of flattening our work as researchers, but remember that it need not be a substitution, but, rather, a supplement.


Learning with Digital Archives

Curarium is, at its core, a platform for learning about and through digital objects and collections. It takes advantage of the unique affordances of digital interaction with a collection — viewing thousands of objects side by side, rapidly sorting large collections, building and sharing thematic “trays” of objects, mapping annotations to specific areas of a digital object’s surface.

Artist Wendy McNaughton writes about the potential of learning through digital interaction in her piece discussing the Metropolitan Museum of Art’s online collection (which I stumbled across through Austin Kleon). McNaughton describes being able to zoom in on a painting close enough to “deconstruct it’s surface anatomy” and gain new insight into technique and process through an exploration at this scale. While this is in no way a replacement for seeing the physical object ( ”[Digitally] You can’t get the Thingness of the painting that is so inherent in the making of any object, of any painting”) the digital representation allows for discovery and learning that would not otherwise be possible.

Under the Hood 02: Database

The development of Curarium is taking place at an interesting time in the history of databases.

For a long time, relational databases (commonly called SQL databases, after the Structured Query Language used to interface with them) have been the dominant choice for storing data in web applications. These databases work as a collection of interconnected tables, think hyperlinked spreadsheets, where data is stored within a predefined structure. Popular examples of this kind of db include MySQL, PostgreSQL, SQLite and Oracle Database.

However, in the last few years non-relational databases (aka NoSQL) have seen an increase in popularity, and trends seem to indicate that they will be dominant in the not so distant future. This group includes divergent approaches to data storage and manipulation, but for now they are mostly confounded by their choice not to adhere to the SQL model. Examples of this type are MongoDB, Redis, Apache Cassandra and Elastic Search.

One of the many benefits of NoSQL databases, is that the data stored can have an arbitrary structure which needs not be predefined. This made them a very attractive choice for Curarium, a platform that seeks to ingest many different types of structured data (museum and library records, API responses, etc.) and make it interoperable. However, traditional SQL db’s still offer many advantages, not the least of them being their level of support, ease of deployment with popular web application frameworks and the SQL language itself.

Luckily for us, PostgreSQL offers solutions that bring to the table the best of both worlds: the JSON and hstore datatypes, which allow us to save the data in its original structure, and to parse into a customizable representation that is Collection specific. JSON (JavaScript Object Notation) is a data exchange format “discovered” by Douglas Crockford in the early 2000′s, which has become the de facto language for RESTful API responses. It is a common, accessible format which means that it is easy for any data contributor to provide material in the JSON format, and it allows for arbitrarily complex hierarchical structures. PostgreSQL has added json as a native type, meaning data can be stored and manipulated directly as JSON. Along with this, Postgres also offers hstore, a schema less key value storing system. The advantage of hstore is, while supporting less complex structures than JSON, it can be indexed to provide faster searching and manipulation. So, Curarium uses JSON to store the original record data, and hstore to store the parsed version that the users interact with, through search results and visualizations, effectively using only two ‘columns’ of a traditional relational database to provide Collection tailored data.

Collaboration in Curarium: Teaching

Curarium is a platform that users can use to work with digital art records in many nuanced and user-friendly ways. But it doesn’t just seek just to enhance a single user’s ability to find, view and manipulate a record. It seeks to enhance the experience of the entire community of users and facilitate the possibilities for discovery and collaboration.

Today we’d like to reach out to teachers, to begin to imagine Curarium as a space for the virtual class room.

1. Create a Circle for your class
2. Group in that circle all of the Collections that are relevant to the class
3. Invite your students to join your Circle

From the Circle Dashboard you can access your collections, you can see the newsfeed of student activity: who has been annotating and what, who has created and shared Social Trays, who has started a Spotlight, who has published a Spotlight, what discussions are spiraling off from these initiating activities.

You can assign work to your students on the Circle Dashboard, you can comment on your students’ work, your students can comment on each others’ work.

You can begin to work collaboratively in a space that goes beyond both the screen and the classroom.


Spotlights and object stories

In Curarium, you can assemble objects, visualizations, text, and annotations into stories called Spotlights. The platform for creating these spotlights from Curarium trays is called Waku, and is currently being developed in collaboration with the Digital Archive of Japan’s 2011 Disaster. (The name Waku comes from the Japanese word for “frame.) It’s built using Javascript, leaning heavily on Node.jsBackbone.js,and Bootstrap, with Kinetic.js for canvas manipulation. With Waku, Curarium users will be able to create multimedia stories with richly annotated objects at their core.

Waku won’t be fully integrated with Curarium until post-beta launch, but we are already playing with Homeless Paintings:

Screen Shot 2014-05-21 at 11.36.56 AM

Under the Hood 01: components

Like all modern web applications, is made up of many different components, designed to work independently (it is, itself, a collection of tools).

At the core of Curarium is the Ruby on Rails application framework, which is the glue that holds all of the parts together, and the brain that coordinates them. This is were collections, records, spotlights, circles, etc, are defined.

All the information stored by Curarium, starting by the thousands of individual items that comprise collections, is stored in a PostgreSQL database. We we wanted to keep the structure of our data as flexible as possible, without compromising the stability and ease of deployment that comes with well tested databases. PostgreSQL meets all this requirements, being one of the oldest open-source SQL databases, and has the added benefit of native JSON storage, which allows us to ingest arbitrarily structure data.

On the front-end, besides the usual suspects like HTML5 and CSS3, we are of course using jQuery, plus a few other task specific javascript libraries:

  • D3 is our main visualization library. It offers a comprehensive set of predefined diagrams and charts, while remaining very customizable.
  • Kinetic.js is the library we are using for manipulating the HTML canvas. It powers our record navigation and annotation tool, allowing us to zoom, pan, and graphically tag the images.
  • Finally, WYSIHTML5 will take care of rich text editing for content creation.

Shortly we will be taking a more in depth look to how some of this components work within Curarium, but each of them is worth checking out on its own!


Today we began thinking about ways to use our developing set of iconography to depict some of the more complex features of Curarium, such as circles. Circles allow users to share and collaborate on work within a specified collection or set of collections. When you explore any collections in your circles, you can use a social tray to work in real time with circle members to collect and order records.