Going forward I will be working with the WCMA dataset. There is a good amount of data, and it will give me good experience with analyzing datasets and figuring out where and how visualization can be used. I have done some previous basic visualizations with data from this dataset, but I hope to dig deeper and see what else I can visualize.
There are a lot of questions to keep in mind; who is the audience? What is the goal of visualization? What do I want out of this visualization? Where might this visualization be displayed? How might it be used? I know that the overall theme I want to work with is browsing: how can someone visually browse the collection, and what might they be looking for? What messages can I find out about the collection as I browse through it myself?
First I looked at how you can currently browse through the WCMA collection. The pieces are categorized by medium, but you can search through the entire collection by various attributes (title, artist, medium, classification, etc). Though you can search by description, the descriptions vary greatly (from lists of adjectives to sentences). Scrolling through the collection can be tedious and overwhelming, since only several pieces are showed at once on a given page, and the size of each thumbnail is different, which means that there isn’t a uniform grid. The pieces can be arranged alphabetically or by some other attributes, such as collection number or date, but I found myself skipping pages or not paying attention to a lot of pieces simply because they blurred together perceptually. Clicking on an individual object is very informative, providing multiple thumbnails and detailed information about the piece. This raises an important issue that I will keep in mind: how do I present an entire collection in a way that isn’t overwhelming?
WCMA currently has a gallery-classroom called Object Lab, where professors of Williams College work with the museum to choose pieces from the collection both to display at the gallery and also to use as class materials. I thought this would be a fruitful area to learn more about, as it will tell me more about how people with a range of different familiarites with the collection browse it with a theme in mind. I talked to Elizabeth Gallerani, who curates Object Lab, as well as professors who have collaborated with Object Lab in their courses.
I learned a lot from these conversations, and my major takeaways are:
My focus is on information visualization, so I will not be doing extensive research in figuring out how to make my visualization more practical, but my research so far has given me a much clearer picture of what to keep in mind as I start brainstorming and thinking about what I want to do. I have a general sense of what people may be looking for, as well as what may be feasible given the constraints of the dataset and time: that is, a way to look through a bunch of different features and also to visually scan the collection in a non-overwhelming fashion. My first intuition is an interactive visualization that highlights certain aspects or certain pieces of the collection. This will entice viewers to focus on and explore highlights that cath their attention. In other words: overview, zoom, filter! My next step is to see what visual features I can extract from the dataset, since that will be very helpful in figuring out what is feasible given the current data. Perhaps as I continue to work on this project, I will encounter interesting insights that I can communicate through visualization. For now, my goal is to develop this browsing-through-visualization idea.