Statistical Graphics Working Group Seminars

  • 4/30/2015 Xiaoyue Cheng "Enabling Interactivity on Displays of Multivariate Time Series and Longitudinal Data" In multivariate time series we expect to see temporal dependence, long term and seasonal trends, and cross-correlations. In longitudinal data we also expect within and between subject dependence. However, there are a few issues in visualizing large temporal and longitudinal data with static plots. Interactive graphics will effectively facilitate exploring the temporal components. In this talk I will introduce a taxonomy of interactions on time series and longitudinal data plots, and describe new methods for data pipelines, data transformations, additivity of interactions, and linking, to construct the specific interactions. The ideas are implemented into cranvastime, a part of the R package cranvas. The package provides many different types of interactive graphics that can be used together to explore data or diagnose a model fit.
  • 4/23/2015 Karsten Maurer "Binning Strategies and Related Loss for Large Data Visualization" The demands of large data require us to re-think strategies of visualizing data. Plots employing binning methods have been suggested in the past as viable alternative to standard plots based on raw data, as the resulting plots tend to not be as affected by increases in data size. We will discuss properties of binning algorithms associated with the construction of binned scatterplots for visualizing bivariate data. Binning algorithms inherently carry loss of information during aggregation, so we quantify this by defined loss functions specific to the deterioration of visual attributes. Binning and loss properties can be explored through simulation and case studies of real data; yielding practical recommendations for specifying binning algorithms used in constructing binned scatterplots.
  • 4/16/2015 Di Cook "The Statistical Eye: Reading Patterns in Visual Displays of Data" Joint work with Omesh Johar, Niladri Roy Chowdhury. Subjects were studied using an eye-tracker to read three different types of statistical plots, containing data that had been simulated to have various features. The choice of plots was motivated by patterns in results observed during some of the Amazon Turk studies of visual inference.
  • 4/9/2015 Paper discussion: Lane Harrison, Fumeng Yang, Steven Franconeri, Remco Chang Ranking Visualizations of Correlation Using Weber's Law
  • 4/2/2015 Sam Tyner "Network Visualization in ggplot2: geom_net" There are many implementations of static network visualization in R, but none are equipped with the flexibility and functionality of ggplot2. geom_net was created to fill this gap. Using two data frames to describe vertex and edge information, geom_net makes use the underlying structure of ggplot2 to visualize networks.
  • 3/12/2015 Vianey Leos "Can't Stop Won't Stop: Sharks on the Move" For animals like sharks, GPS trackers and accelerometers are vital to knowing where they go and how they move. Some sharks can be tagged and followed, while others are tagged with more advanced equipment that can produce observations at very fine time resolutions with high accuracy, quickly leading to data set sizes in the millions. I will present various shark datasets, approaches in animal movement modeling, and statistical and biological challenges of analyzing massive amounts of movement data.
  • 3/5/2015 Jennifer Chang Mango: an integrated environment for network visualization and exploration Mango is a graph analysis and visualization software designed to handle many large graphs with tens of thousands of nodes and millions of links at once. Mango combines the power and flexibility of a high level Graph Exploration Language (Gel) with the easy operation of a graphical user interface. The result is a general purpose graph analysis tool that can be used to analyze many heterogeneous biological data sets. Mango is not limited to handling only biological data - Any data association and linkage information can be loaded into Mango and analyzed.
  • 2/12/2015 Discussion lead by Heike Hofmann "Perceptual kernels" paper 1 , paper 2 , paper 3
  • 2/5/2015 Two short seminars of Stat 579 (Fall 2014) projects: "Ebola Now and Then: What is different about the 2014 West Africa Outbreak?" Jun Fang, Jing Zhao, Rafael Martinez-Feria, and Carrie Chennault; "Estimate the trend of social atmosphere in the US by examining popular songs played over" Taikgun Song, Hyeongseon (Sammy) Jeon, and Sanghoon Cho
  • 1/29/2015 Discussion lead by Carson Sievert, read the paper "Graphical Perception: Theory, Experimentation, and Application to the Development of Graphical Methods" by Cleveland and McGill.
  • 1/22/2015 Discussion lead by Sam Tyner Infovis and Statistical Graphics: Different Goals, Different Looks by Andrew Gelman and Antony Unwin. Two reactions to the above article by Hadley Wickham and Robert Kosara. Also, check out link for two more reactions by Stephen Few and Paul Murrell, and Gelman and Unwin's response to all the responses. What is Visual Analytics? Read pages 1-18, the executive summary of the goals of R & D for visual analytics and What's an Infographic?
  • 1/15/2015 Discussion lead by Di Cook When you have some time this week, before Thursday, go over to the Food Science building and look in the courtyard. There are a set of sculptures produced by Chuck Ginnever on loan until July. There are 15 identical shapes, but they must be arranged with each one oriented differently from all others. Its a little like the paper-folding tests we have all done. The museum staff need some math-minded people to help them configure the sculptures, and also suggest challenges that could be provided to students, through the iastate web site, throughout the life of the exhibit.
  • Fall 2014 seminars