There's a lot more stress and emphasis on how we communicate science to the general public. The philosophy I'm trying to show students is to become better communicators of their process and to become a part of the reproducible-science movement.
Dr. Tyler Davis Pioneers a Multifaceted Class in Data Reproducibility and Problem Solving
• With only 5 days before class would begin in the summer of 2020, Dr. Tyler Davis used cloud-based technology with ArcGIS to create an asynchronous remote learning environment for his students.
• ArcGIS Notebooks helped to provide accessibility for students in Dr. Davis’s class that needed to meet requirements for almost every degree offered at William & Mary. The program also helped each student to push toward the next step in their level of programming knowledge.
• Since the credibility of scientific data is being challenged now more than ever, Dr. Davis created an environment conducive to creating reproducible work. With that ability, students are given practical experience they can easily share with potential employers.
With the onset of so many changes right before the 2020 summer semester at William & Mary (W&M) in Williamsburg, Virginia, Dr. Tyler Davis decided to truly lean into his calling as a solver of complex problems. Just days before teaching was to begin, he completely redesigned the structure of his Advanced GIS Analysis and Programming class.
Davis's class is part of the Center for Geospatial Analysis (CGA) at W&M and functions as a campus resource—rather than a major or a department—for all things geospatial. The center plays an educational role at the undergraduate and graduate levels and offers technology and geographic information system (GIS) technology and support to faculty, staff, and the greater community. "What drew me to the CGA was that we help solve other people's problems,” Davis said. “That resonated with me because that's essentially who I am. As an engineer by training, I like to solve problems."
Students with different skill levels, career paths, and intentions enroll in Davis's class to increase their understanding of programmatic approaches to spatial data and analysis. Consequently, Davis needed to structure his class to accommodate every student's unique need. What's more, he had to overcome these hurdles amid uncertain times, equitably and in a completely digital setting.
When Davis learned about ArcGIS Notebooks, he realized he had found a key piece that links the framework for showing GIS as a workflow. "A lot of problems can't be solved just by clicking a button," he said. "We have to put different pieces together."
His overarching goal was to help his students understand that solving complex problems—like those they will eventually confront in their workplaces—requires a lot of data, analysis, collaboration, and hard work.
Notebooks is designed to help users better collaborate with other analysts and data scientists, a concept that Davis wanted to carry into the foundation of his class. ArcGIS Notebooks is built on top of the Jupyter Notebook, an open-source web application that allows users to create and share documents that contain live Python code, visualizations, and narrative text. As a result, students can share their work as well as build on work that has already been created by others.
Adapting to Asynchronous Remote Learning
One of the biggest problems Davis planned to overcome was the students' diverse schedules and the potential need to quarantine at unpredictable times. With not much time to prepare, the class adopted an asynchronous remote learning approach, which allows students to view lectures and do their work anytime within their own schedule. Given the complex nature of the class subject matter, this was no easy feat.
"We tried to fix that with a virtual desktop, where they would log in to computers remotely," Davis said. "There was internet lag, and just the earlier versions of the ArcGIS Desktop software we were using was really slow. When you're teaching something that's already frustrating, adding another layer of frustration is not really what you want, because students will likely give up."
When Davis's class moved completely online, he wanted to provide students with equitable computing resources to run GIS software. Davis said, "The ability to put everything in the cloud meant that there was more equality in terms of students having access to all of the tools that I felt we needed. We have trouble with Mac users, both for accessing software and using it. Opening up both ArcGIS and ArcPy in one platform in the cloud using ArcGIS Online, I think, really rang a lot of bells for me," Davis said. Right now, Davis is working on building up their external enterprise server to improve data access to their constituents.
The Value of Teaching Reproducibility
With political tensions high and public skepticism of science increased, Davis realized his students should focus on careful processes that could be easily reproduced for testing and verification. By combining Python scripting with interactive visuals and descriptions, Notebooks allows students, researchers, and educators to explain their analyses in a dynamic, shareable workspace.
"There's a lot more stress and emphasis on how we communicate science to the general public," Davis said. "The philosophy that I'm trying to show students is to become better communicators of their process and to become a part of the reproducible-science movement."
Part of the class requires students to choose a published paper and grade it on the project's reproducibility. In doing so, they see how others have put together projects, and they come to understand advanced geospatial techniques along the way. Students are encouraged to work together on these projects so that they can evaluate how others approach problem-solving with varying levels of knowledge and experience.
Creating a Workflow Conducive to Problem-Solving
Notebooks is designed to ensure that users can better collaborate with other analysts and data scientists. With this software implemented in Davis's classroom, students gain relevant experience with sharing their work and building on work that has already been created by others. "Notebooks can be a part of the production line for communicating science through Esri's ecosystem of applications," Davis said. “All of these apps work well together.”
When students solve problems with time and scale considerations, they can take advantage of programming to meet those challenges as well. For example, when they are geocoding addresses, programmatic approaches speed up the time it takes to scrub the addresses from a website and preprocess them into an acceptable format for geocoding purposes.
Davis’s class also uses Notebooks to look at the delineation of watersheds given a pour point location along a given stream or river network. This requires linking several processes together, which is tedious in ArcGIS desktop and ArcGIS Pro. Putting this process together in a model with ModelBuilder or Notebooks allows the process to be automated and easily scaled to delineating dozens of watershed areas in a fraction of the time.
Of all the benefits that Notebooks brought to Davis's class, perhaps the most useful is that the software enables students to show their transferable skills to a potential employer. Davis provides students with nine different Notebooks tutorials on foundational topics including programming basics. Students finish the class with an understanding of how and why programming is used as well as the ability to visually show their work, which can help them acquire an internship or job.
For the future of the Center for Geospatial Analysis, Davis envisions a community where mapping and spatial analysis are essentially ubiquitous in teaching and learning, research, and campus support. His multifaceted class, backed by supportive tools, has proved to be a significant step forward.