Using the Project EDDIE Remote Sensing of Plants and Topography in R Module in GEO 424: Advanced Remote Sensing
About this Course
GEO 424: Advanced Remote Sensing
EDDIE Modules(s) Adopted and/or Adapted
This module uses big data methods to consider environmental patterns from local to continental scales.
Relationship of EDDIE Module(s) to my Course
This exercise was implemented at the end of an advanced remote sensing course, where students have already been introduced to many of the concepts as well as programming in R. Most of the course is focused on satellite remote sensing, so here students worked with airborne data (higher spatial resolution) for the first time, and this was their first time using R to look at raster imagery. The lectures associated with this lab focused on the technical details of imaging spectroscopy and lidar. Due to the COVID-19 pandemic, this lab happened online (students had remote access to a computer server to access data). Since many students could not run Zoom while being remotely connected to the server, this lab happened mostly asynchronously. As such, we provided them with the R code up front. In an ideal situation the instructor would type the code in real time with the students in a classroom, in a 'Software Carpentry' type of approach, and for part C students would be able to discuss together their preferences. We used an online forum to discuss options, but it was not as active as an actual discussion might have been.
What key suggestions would you give to a colleague before they used the activity in their teaching?
I think this exercise could be useful from a number of different angles - teaching R, thinking about local vs continental scale questions, thinking about remote sensing, or thinking about plant assembly theory. That said, trying to cover all of those different angles might be too much, so focusing on one or two would make more sense. For example, my class is a remote sensing class without an explicit focus on plant ecology, so I mainly emphasize the coding and remote sensing angles, not the plants.
How did you address challenges in teaching with the module?
The main challenge was doing this online when I had planned to do it in person. Using a discussion forum to decide what patterns to look at for part C worked ok, as did having students email the TA their results for Part C, but I think it would have been more fun to do it more interactively.
Students seemed surprised at how messy the data were. Seeing that what looks like a random cloud of points can have a Pearson's R of 0.4, and that those relationships are considered meaningful, I think was a surprise to many of the students. Most students are used to seeing maybe 10 data points on a plot, so seeing 10,000 was new.
This module was the third of three R-based labs in my class. Several students had never used R before, so this was all new. Some students found R much easier than using a spreadsheet program, which was a nice surprise.