2022 Lightning Talks

Ben McCary

Persuasion with Pretty Pictures: Using Data Viz to Make a Point

R is an amazing tool for creating world class data visualizations. Using data wrangling, the tidy verse, and a lot of rummaging through stack overflow you can do just about anything…even if the point of the viz is just to impress people into listening to the rest of your talk! In this talk, see how to use themes, functions, and a reproducible data pipelines to create impressive data visualizations.

Brittany Barker

Oregon State University, Corvallis, OR

Spatial Modeling in R to Help Detect Emerald Ash Borer, a New Invader in the Pacific Northwest

Emerald ash borer (EAB), Agrilus planipennis, is considered to be the most destructive and costliest forest pest in North America, with populations continuing to invade rural and urban areas throughout the continent. Millions of ash trees across 36 U.S. states and five Canadian provinces have been completely decimated since the wood-boring beetle was discovered in Michigan in 2002. In July 2022, EAB was discovered for the first time on the West coast in a park in Forest Grove, OR, about 25 miles west of Portland.

I will present on a spatial modeling tool written in R that can help with efforts to detect EAB early, before it has time to establish and spread. The tool combines gridded climate data with species-specific physiological information to produce forecasts of climatic suitability and the timing of life cycle events across North America. For example, model forecasts of adult emergence can help ensure that traps are installed well before adult beetles begin exiting trees in the spring and summer. Forecasts of climatic suitability can provide insights into the potential range limits for EAB, such as in very cold parts of Canada. Model forecasts and the R tool are freely available at https://uspest.org/CAPS. Our new grant will fund work to make forecasts more interactive, increase the number of end-users, and engage citizen scientists in contributing observations for forecast validation.

Kevin Kuo

Hal9, Seattle, WA

Bussin: A Modern Framework for Interactive Data Apps

Feel like shiny is outdated, too restrictive without hacking javascript, or just too hard to learn? Check out bussin, a framework for building web apps with an intuitive execution model and simple visual frontend designer.

Meike Niederhausen

Oregon Health & Science University, Portland, OR

Using visualizations to understand complex care trajectory trends of Veterans in Home-Based Primary Care

Home-Based Primary Care (HBPC) is a home care program for medically complex Veterans, providing comprehensive primary care from an interdisciplinary team of Veterans Health Administration (VA) staff. With a rapidly aging VA population, the VA will need to provide more long-term care services and support to care for Veterans. A goal of HBPC is to increase patients’ time at home and allow them to age in place. We analyzed daily longitudinal data over two years of 11,640 Veterans in 440 different HBPC programs across the United States to better understand transitions between different care type settings, such as home, emergency department, hospital, long term care, and hospice. Visualizations played a key role in communicating the complexity of the care setting trajectories. For example, we adapted heatmaps, raster plots, and alluvial plots to display longitudinal data of many patients simultaneously and identify trends in care. These were created using the ggplot2 and ggalluvial R packages.

Siddharth Bhatia

ProCogia, Vancouver, BC

Engineering Better Shiny Apps Using Docker

More than often, there are differences in dev environments across the team for the same app in development.Additionally, there is a perception that Shiny is better for POCs and not so much for professional web applications at scale. In this talk, audience will have a chance to learn how to make Shiny app development a part of a more professional software engineering process. They will learn how to synchronize their dev environments, locally test their apps (be it unit testing, load testing or user testing), and be part of a CI/CD system using Docker and Shiny with some help from Github Actions and Shinyapps.io. Some technologies that help scaling dockerized apps will be also be hinted as an exploratory task for suitable audience.