Assumptions, Visualizations, and the Quantified Self

Friday February 6 at 10 AM, the final session of DST4L’s third round commenced. Jim Davenport had more than enough content to cover during his marathon session. After a brief introduction to his background, his second life as an astronomer, and his blog ifweassume.com, Jim explained the methods he relies on to find stories worth sharing. Sometimes he starts with a question and looks for data to answer it, other times he lets the data lead him to a story. He illustrated both approaches with a variety of examples from ifweassume before discussing the importance of good design and the quantified self. Jim originally used IDL for many of the projects he showed the class but he prepared some of his work as iPython notebooks so participants could get hand-on experience making the visualizations themselves. As the day wrapped up, members of the class shared their feelings on the role of the library in data driven research.

Jim’s introduction to his astronomy background provided the class with an understanding of his perspective of the world and his approach to information visualization. Jim illustrated the importance of good visualizations in the sciences and transitioned to the content from his blog by sharing a fascinating video he created as part of his PhD research.

A post explaining the video is here: http://www.ifweassume.com/2015/01/12-days-of-data-from-kepler-17.html

There are multiple paths to answering a question and often many solutions as well. Maps and graphs make both the question and its answer easier to digest. Jim showed the class several examples of his approach to answering new questions that might not have existing data. One such question was, “How far is the nearest coffee on University of Washington’s campus?”.

UWCoffee

This data then led him to investigate Starbucks locations across the US.

Starbucks 

When taking the opposite approach and searching for interesting questions in known datasets, a good visualization can reveal questions and their answers simultaneously.

airports  DEWLine

They can also lead readers to discover a topic on their own terms. A person tends to be more engaged with something that they must be actively invested in.

artDimensionsPlot

Jim made sure to stress that data presented without regard for aesthetics can be unpalatable. Attention to the use of color and emphasis is valuable when designing for a particular purpose. It is important to balance the need to display information with a viewer’s desire to see something visually appealing. Sometimes the trade off leads to a more pleasing image with less information density.

artDimensionsOverlay

Some artwork emulates the display of data.

artOrData

Jim closed with a discussion of the quantified self. Benjamin Franklin and Stephen Wolfram are noteworthy figures who both exemplify the notion of measuring one’s life. Besides being interesting, the data can be used to notice and change behavior.

wolfram

It also creates benchmarks for various kinds of success.

websiteViews

batteryLife

While ifweassume tends to be focused on data visualization, its heart lies in Jim’s commitment to discovery and creation. Jim’s work is inspiring not only because of the results, but because of his resolute pursuit of answers to new questions. His final products serve as markers for the kind of work he’s doing. It is a worthy ambition to make graphs like Jim’s, it is worthier still to seek knowledge in the hands-on way that he does. These ambitions are the building blocks of science and human advancement.

All images in this post are from Jim’s slides. To view his slideshows, please visit: https://speakerdeck.com/jradavenport

Jim’s Github repo for the day is available here: https://github.com/jradavenport/dst4l_2015

Megan Potter’s notes from the day are available here: QuantifiedSelfJimDavenport

Leave a Reply