The Statistical Atlases of 1870 to 1890 are wonderful sources of statistical graphics, created in a time when conventions for statistical graphics were still very much in flux. We will be looking at some examples of visualizations from the 1870s - some of which will look very familiar, discuss the charts from their cognitive perspective, and look into re-designs based on available Census data.
There will be cake - at the point of writing, the baker is not certain about pie versus box shape.
Read more →
I will talk about visual hypothesis testing using lineups, including how we model the process of choosing a plot from a lineup, different types of lineups, and how we estimate a lineup’s difficulty. There will also be a lot of pictures of puppies!
The slides can be found here.
Read more →
While summer may be a time to relax and enjoy the sun, it is also a time to enjoy working on research without the stress of classes (possibly while also sitting outside and basking in the sun). In this talk, returning graphics group participants will tell about summer research projects, conferences and workshops attended, and internship experiences. We hope this will allow everyone to learn about the work done by others, provide some inspiration going into the fall semester, and allow new attendees to get an idea of the breadth of work done by individuals in the graphics group.
Read more →
One of the first things we need when we build statistical models is a rich set of training data. Creating a labeled training dataset is feasible when the dataset has hundreds, or maybe even thousands of rows. But the top-performing models on many machine learning tasks do best with millions or even billions of examples. How do we automatically build a large training set with billions of examples? Enter BERT, Google’s new Natural Language Processing (NLP) language model.
Read more →
One of the more recognizable terms associated with analysis in sports is “sabermetrics”. Made famous in part by Bill James in the 1970s, and executed with some success by Billie Beane and the 2001 Oakland Athletics, it has changed the way people view sports. Sports are no longer just about intuition. Following this paradigm shift, the use of analytics has grown tremendously within Baseball, and has even made its way into other sports, most prominently with Football and Basketball.
Read more →
