The final project for MAT259: Algorithmic Data Visualization involved visualizing data in 3D.

Since we are using Library data I chose to investigate the Dewey Decimal category information. The Dewey Decimal system places every item in one of 10 broad categories that each have many subcategories.

I wanted to investigate how the distribution of Dewey Categories varies over time so I plotted each day as a column with height proportional to the number of transactions. Each column is made of color-coded solids whose heights are determined by what percentage of daily  transactions each category represents.

Each category is connected with a line to emphasize how the trend varies over time. The solids and lines can be independently toggled to vary the displayed information.

Clicking a column will display similar distribution data for the 10 subcategories within each main category — 100, 110, 120, etc. Here the width of the solid is proportional to its relative percentage within its category.

The applet displays data for the entire year of 2008. The month can be changed using the ‘[' and ']‘ keys or by clicking “Next Month” or “Previous Month” on the calendar near the month name.

The visualization can be rotated by clicking and dragging the mouse. Additional camera controls are explained on the applet page.

The second project for MAT259: Data Visualization involved exploring two dimensional space for visualizing data from the Seattle Public Library. I was interested in getting away from the traditional chart mentality and displaying data in a more unique way.

I settled on comparing two numerical quantities using a seesaw or balance/scale metaphor. I chose to enumerate the number of checkins and checkouts over time to analyze how they relate and how the balance changes over time.

I found that most days the trend is about flat except for in the first part of the day when the library is not open but people are able to perform certain functions through the library website.

(more…)

The first project for MAT259: Data Visualization involved exploring ways to display library checkout data in a 2D space. I chose to investigate how long each item was checked out and how it relates to item type, category or shelving section. 

My goal was to be able to make qualitative judgements on how checkout duration trends vary over various item categorizations. For example, are Children’s items checked out longer than Young Adult items? Are Books checked out longer than DVDs?

Click the image below to try the applet out and note the keyboard options described below the program.

The demonstration program incorporates only one hour of checkout data to reduce loading time.

Below is my concept sketch and some screenshots from the program.

Visualizing Trends in Item Collection, with sidebar controls

Visualizing trends in checkout duration using alternate method of allocating x-coordinates

The image above uses a special method for assigning x-coordinates that is based soley on the how the checkout duration compares to the maximum duration in the dataset. Because it does not depend on the position of the transaction in the dataset it is unaffected by sorting parameters. This view is useful to quickly see where checkout durations congregate (brighter green areas) or where there are no transactions with durations in a certain range (gaps).

To keep the visualization from getting too cluttered, catagories representing less than 2% of the dataset have been ignored.