September 21 and 28: Recap & Next Steps
This is a two-week recap covering Session 4 and 5 along with the revisions to the course we discussed the last couple weeks. Looking forward to seeing you all online soon!
What did we do for Session 4
Introduction to filtering data with QGIS
We spent most of our time on September 21 reviewing how to work with QGIS including:
- How to load data from a GeoPackage file or a CSV file with coordinates (also known as a delimited text tile)
- How to filter data using the point-and-click GUI and using basic QGIS expressions
- How to export filtered data to a new layer and a new file
I demonstrated some of the most common ways mistakes (mostly inadvertently) so make sure to remember:
- When selecting features in the Map view, you must have the layer of interest selected in the Layer panel
- When building QGIS expressions, character literals must be wrapped in single quotes; double-quotes are reserved for field references
We primarily focused on how to use the “select by expression” interface but you can find options to use expressions all throughout QGIS including:
- Custom labels (e.g. concatenating multiple fields with the
||operator) - Data-driven layer feature styles
- Field calculator (for changing the value of an existing or new field)
- Using variables in the Layout builder (reference variables using
@and eenclose expressions in[%and%]to evaluate)
For more information, you can review the documentation on QGIS expressions or explore a list of available functions, operators and variables. Many functions in QGIS expressions are also based on the conventions of SQL (short for Structured Query Language) and PostGIS so other relevant resources include the PostGIS reference documentation.
After the QGIS demonstration, we had a constructive discussion about how the course is going so far and came up with a loose plan to extend the first section of the course so we can spend more time during class with hands-on activities. I also agreed to rework the remaining assignments to give everyone more opportunities to practice the skills described in Lovelace, Nowosad, and Muenchow (2022).
We closed the session with me attempting to demonstrate a workflow for summarizing data using my Baltimore Tax Sale data as an example. The code for that analysis turned out to be less immediately reproducible than I remembered so we’ll return to this topic with a more straightforward in a future class!
What did we do for Session 5
We dedicated most of last week’s session to a presentation and discussion with Reina Chano Murray, Geospatial Data Curator and Applications Administrator, Johns Hopkins Sheridan Libraries. Reina graciously shared her slides and a collection of related resources in a GitHub repository. Here Reina’s top tips for documenting your data:
- READMEs are your best friend
- Document your data along the way - saves you time at the end
- Use descriptive file names
- If you’re using geospatial desktop software or web GIS, create your metadata in the platform/software you start in
Questions
We also caught up on a few questions from the last couple weeks of log entries:
Is ggplot typically used more for exploratory work or is it also a good way to make “publication-worthy” maps and visualizations?
It is definitely a good way to make “publication-worthy” maps and plots! There are many ggplot2 extension packages. Among the 100+ packages are {ggpubr} (specifically designed to create publication ready plots) and {hrbrthemes} (a popular collection of typographic-centric themes). I also shared a simple example showing how a recent map from the Baltimore Banner could be effectively replicated by using system fonts and a carefully tweaked theme. Check out this post on Setting up and debugging custom fonts by June Choe for more details on using system fonts.
Is an inner join the same as a right join?
Nope. dplyr::inner_join() and dplyr::right_join() are both examples of what Wickham and Grolemund (2022) describe as mutating joins. These functions typically have two parameters (x on the left and y on the right). A left join keeps all observations in x, a right join keeps all observations in y, and an inner join keeps only those observations found in both x and y.
How can you “merge” two different data sets when it doesn’t make sense to use dplyr::left_join()?
You can combine sf data frames by using dplyr::bind_rows() to stack one data frame on top of another. You can also use the base R rbind() function but only if the two dataframes have identical columns. dplyr::bind_rows() is more flexible and just adds new columns for any attribute that appears in the second data frame but not the first.
What to expect for Session 6
Due to the schedule conflict with Yom Kippur, our next course session is this Monday evening (tomorrow!). We’re diving back into R to review how to use sf for spatial operations (functions dealing with spatial relationships between features) and geometric operations (functions that modify the geometry of features) and, hopefully, start talking about how you can use these functions as part of an exploratory analysis.
Updates to the course schedule
Right now, you can compare the schedule from the start of the semester with this updated schedule. Once we finalize these changes, I’ll replace the schedule with the updated version. Here are the highlights of the changes:
- I updated the titles and descriptions for the last five weeks to more accurately reflect the materials we have been able to cover so far
- I added two new sessions to the first half of the course: Tidying and summarizing data in R and Summarizing and analyzing data with R (extending this first half by two weeks)
- I added a new session on Editing and creating geometry with R and QGIS to the start of the second half of the course and reworked the descriptions for Getting data from public web services and Working on collaborative data projects
- I temporarily removed the assignment due dates while I rework the remaining assignments to function more effectively as exercises.
We’ll briefly review these changes during our October 3 session and I’ll share an updated activity schedule and more details on a final project ASAP.