teaching an old dog new tricks: learning QGIS


In my last post, I showed that my lower level undergraduate students were performing quite well in their online transition, when learning quantitative geography. In this post, I want to focus on my Advanced GIS students. As you can see from the syllabus, this is an advanced course, focused on ArcGIS Pro, and many of the extensions such as Spatial Analysis, Geostatistical Analyst, Model Builder, and even some Arcpy. But, like everyone in the country, everything went haywire around mid March.


Fortunately, I had gotten my students through all of the ArcGIS Pro tools, and we were transitioning to final projects. I knew that it would be impossible for them to conduct sophisticated final projects while being locked down at home, so I decided to have my students learn QGIS Desktop. To do that, they would follow along with my online QGIS Desktop course as part of the Learning the FOSS4g Stack curriculum. Similar to my Quantitative Geography course, I opened the learning up to universities around the world.


Having the students make this transition was opportune: many people are exploring the use of open source GIS, and one might wonder whether it is possible to learn a new GIS product on your own (with instructional videos, of course) in a matter of weeks, and then become productive in its use. I know many have said I'd like to learn QGIS, but I fear the learning curve will just be too much. Is that true? Well, now was my time to find out.


One of the tasks I had the students do was to recreate our first ArcGIS Pro lab three different ways: using the QGIS Desktop GUI, using the model building capabilities in QGIS, and finally using Python with ArcGIS. They had already performed the lab in ArcGIS Pro and Esri's Model Builder during the face-to-face GIS class, so now was an opportunity to see if they could complete the labs with other software they hadn't used before.


The lab exercise


This is a lab I've been using for over a decade. As students move into Advanced GIS, the first lab is basically designed to make sure they have a basic understanding of GIS tasks that they should have learned in Introduction to GIS. Nothing special, just a lot of repetitive use of Merge, Erase, Select By Attribute, Select By Location, Summarize, Buffer, Intersect, Join, etc, all under the guise of finding an ideal location for an ecological site.


The students used a lab exercise, written for ArcGIS Pro. That worked out very nicely, as they knew what commands to use in order to complete the tasks. However, when we transitioned to QGIS Desktop, I did not change the lab directions. Instead, I told them that part of the exercise (and the measure of a good GIS analyst) is to take generic directions and learn to integrate them into any product.


I have 13 students in this class, and they are all doing the work. But, I asked one student if he wouldn't mind my showing his results. His name is Brad Haas, and he is a senior. Brad is a great student, and also one of our scholar athletes, as a member of our SU baseball team. Unfortunately, Brad did not get to play his senior year as COVID-19 knocked out his season. But, when I was at a game last year, I took this photo of him coming up to bat:




No kidding: he actually hit a home run! Sadly, I wasn't able to get that on my phone! I sent him the photo last year as a reminder of his great work on and off the field. Anyway, in addition to baseball, he's pretty good at GIS, too.


Desktop GUI


The ArcGIS lab probably had close to 30 steps involved between merging counties, towns, roads, selecting land cover, buffering land cover, intersecting the layers, and then summarizing geometric areas. So, it was fairly tedious, but also quite challenging.


Here is an example of an intersection that Brad performed, as shown in his log file:



When using QGIS, there were challenges for sure. But Brad, and all the other students were able to complete the lab and get a final answer. There were a few differences in the results, but only minor, and likely due to the criteria for selecting data, or a minor difference in fuzzy tolerances. But, all in all, the students were able to replicate the original lab using the QGIS desktop GUI.


Model Builder

The students love Model Builder. If you've ever taught your students traditional GUI-based GIS, and then shown them model builder, they treat you like a hero! I constantly get asked why didn't you show us model builder last year?. My answer of course is we want you to understand how GIS works before we introduce you to tools that make the process easier, otherwise, you'll never learn what commands to use.


So, one of our earlier labs is to recreate that first ecological lab using Model Builder in ArcGIS Pro. Here is an example of Brad's first lab, shown using Model Builder:



Efficient, aesthetically pleasing, and very easy to do. At this point, the students were then required to recreate the lab using the model building software in QGIS. Again, here is what Brad produced:


Uglier for sure, but impressive that this college student could recreate his entire multi-step GIS lab in the form of a model in QGIS. And, all of this was done in his home under quarantine without me being there. He just watched a short video that illustrated how to complete the task.


ArcPy


We now have 2 weeks left in the semester, and I have students watching videos on how to use ArcPy within ArcGIS Pro. I teach a GIS Programming course in the Fall, but also feel it is important for students to learn some Arcpy, even if they can't take my GIS Programming class. So, I usually have a 1 week introduction to Arcpy lab. So, for now, I turned to online videos to teach the methods, and will have them recreate that first lab again, only this time using Python. This will give the students a great breadth of understanding: completing the same lab with ArcGIS Pro, QGIS Desktop, ArcGIS Model Building, the QGIS Modeler, and also with ArcPy. I'll post the results next week.


The Takeaway


I'm always very proud of the Salisbury University Geography students. They are hard working, capable, and eager to always learn more. By stepping up and learning new things, they were able to demonstrate that yes, it is not that difficult to learn a new software product in a relatively short amount of time, if you have the right teaching tools, and also the determination to work hard.


So, while you are stuck at home under quarantine, do your professional career a favor and start learning some new software - there's nothing else going on!!


I'll leave it to my students to share what they learned in the discussion below...



413 views

© 2020, Arthur J. Lembo, Jr.