Talks and Tutorials

Data Visualization with Altair: a grammar of graphics for Python

Event: PyData Madison Meetup
Date: 01/23/2020
Abstract: Altair provides an elegant and consistent API for statistical graphics. Altair is built on top of the Vega-Lite high-level grammar for interactive graphics which is based on the “grammar of graphics” idea proposed by Leland Wilkinson. Altair’s key strength is the provision of a clear mental model based on a set of graphical primitives and carefully designed combinatoric rules, that yields an ample space of graphical displays, avoiding the constraints of chart taxonomies. In this talk/tutorial, we will learn the fundamental building blocks of Altair/Vega-Lite’s interface and design.
1. Two approaches to data visualization APIs: tell me how and tell me what
2. Wilkinson’s grammar of graphics
3. Altair, Vega-Lite, and Vega
4. Altair in Practice: guided live tutorial

Introducing software development practices and tools for research in the behavioral and social sciences

Event: UW - Madison’s Data Science Research Bazaar
Date: 01/25/2020
Abstract: Research in the behavioral and social sciences (B&SS) is increasingly relying on complex computational procedures. Nonetheless, researchers in the B&SS usually have little formal training in software development in the context of scientific computing. This situation limits researchers’ ability to produce data processing pipelines that are reproducible, reusable, reliable, maintainable, extensible, and shareable with the wider scientific community. Introducing a set of practices and tools from software development can significantly help to alleviate this situation and improve the long-term sustainability of research that relies on heavy computation.

In this talk, I provide a selection of practices and tools requiring relatively low effort in exchange of high impact on improving researchers’ computational work-flows. I also provide a minimal example illustrating the application of these simple principles in an end-to-end data analysis project.
1. Creating a simple and well-organized data file system
2. Using virtual environments
3. Using version control systems
4. Example 1: Writing a basic reproducible script
5. Example 2: Setting up a machine learning experiment tracking
6. Testing your code
7. Summary and conclusions
8. Resources to learn more