Example

Let’s walk through an example where we will merge external data to the shapefile and then map all the districts in California. The external data is from Josh’s covidedu project, which scrapes district websites for specific words. In this case, the search terms were “covid*”, “coron*”, and “closure”. I highly recommend using covidedu for easy scraping from a lot of district websites!

First, let’s call our libraries.

library(tidyverse)
# if you haven't installed the package yet
# devtools::install_github("ivelasq/leaidr")
library(leaidr)
library(maptools)
library(viridis)
# if you don't have this downloaded
# install.packages("mapproj")

Time to get your data! Use {leaidr} to download and prep your shapefiles for California (FIPS == 06). Read in the external data (in this case, summary-of-table-of-links.csv).

Note: You must have a GitHub PAT set to run lea_get(). Happy git with R has a great walkthrough on how to do that.

# download the shapefile into a designated folder
leaidr::lea_get(path = "./test")
# prep the shapefile for the state(s) you'd like
ca_shapefile <-
leaidr::lea_prep(path = "./test", fips = "06")
# read in the external data that also has NCES ID's
# this is from the covidedu project
ca_data <-
read_csv("https://raw.githubusercontent.com/making-data-science-count/covidedu/master/output/2020-04-29/summary-of-table-of-links.csv")

Join the CSV to the shapefile.

ca_merge <-
sp::merge(ca_shapefile, ca_data, by.x = "GEOID", by.y = "nces_id")

Now ‘fortify’ the data - this converts the polygons into points. This is so ggplot can create the plot.

If you get the error isTRUE(gpclibPermitStatus()) is not TRUE, then you need to enable gpclib by running gpclibPermit() (this is part of the {maptools} package, which should have been loaded above). Note that support for gpclib will be withdrawn from maptools at the next major release, so you might have to try something else if the package has been upgraded.

If you run gpclibPermit() and you keep getting FALSE, then you are missing the package {gpclib}. Install the package, then run gpclibPermit() to set it to TRUE.

(I don’t know if this is the best/only way to do this - if anybody has suggestions, please let me know!)

# install.package("gpclib")
gpclibPermit()
ca_points <- fortify(ca_merge, region = "GEOID")

Now, join the points and the shapefile data.

ca_df <- left_join(ca_merge@data, ca_points, by = c("GEOID" = "id"))

We can finally plot the shapefile and its data!

ca_map <-
ca_df %>%
ggplot() +
geom_polygon(aes(x = long,
y = lat,
group = group,
fill = any_link_found),
color = "gray",
size = .2) +
theme_void() +
scale_fill_viridis(option = "cividis", discrete = TRUE) +
ggtitle("COVID-Related Links Found on CA School District Sites")

To make a nicer looking map, then you can use coord_map().

map_projected <- ca_map +
coord_map()
print(map_projected)

Tada! A full school district map for California.

Call for Collaboration

Please try out {leaidr}! I hope that it is useful to you in your work. I’d love any collaborators to join me in making it easier/better!

• Other functionalities: Thinking of: being able to filter shapefiles by NCES IDs as well as states; adding commonly used data (like district demographics).
• Issues: If you run into any issues, please post on the GitHub page!

Resources

• Joining Spatial Data
• Analyzing U.S. Census Data Using R