Multi-Layered Network of Map Collections at Seattle Public Library
MAT 259, 2019
Susan Burtner
Concept
This project builds upon my last and creates a multi-layered network where layers represent different map collections at Seattle Public Library. Within each collection and across collections, nodes are represented by titles and an edge is drawn for titles that have the same subject. The goal for the visualization is that it communicates which titles are grouped together by subjects within and across collections.
Query
USE spl_2016;
SELECT DISTINCT title, subject, collectionCode
FROM title, subject, itemType, itemToBib, collectionCode
WHERE itemType.itemtype LIKE '%map' AND
collectionCode IN ("camap", "canf", "camapr") AND
subject.subject != "" AND
collectionCode.itemNumber = itemType.itemNumber AND
itemType.itemNumber = itemToBib.itemNumber AND
itemToBib.bibNumber = subject.bibNumber AND
subject.bibNumber = title.bibNumber
ORDER BY collectionCode ASC;
Preliminary sketches
I very much wanted to extend my second assignment into a 3-dimensional representation of networks, so in my initial sketch, you can see how I simply turned my 2D project into an idea in 3 dimensions, with the added connections between collection layers. I also dabbled with looking at other variables and ways to look at these other variables, but opted for what can be seen in my final images.
Process
I did not take many progress screenshots, mainly because there was not much to show until I got the whole thing working. At this stage of my work, I show the top 3 (by size) map collections in the Seattle Public Library in their own layer, but also show how different collections contain similar subjects across collections as well. The nodes are scattered randomly this time, and are represented by simple spheres. The color scheme at this point isn't great, and is modified in my final iteration.
Final result
I received some great feedback from the class, and changed my colors, made the interconnecting edges have a gradient color, and added the capability to hover over a node to see the title of the subject. While I am not the biggest fan of the nods being placed randomly since you can see the edges of the layer so clearly, I think the visualization still does a good job of communicating shared attributes across multiple dimensions.
Code