In MattjCamp/wicher: High school graduate forecasts from The Western Interstate Commission for Higher Education (WICHE) for R

Introduction

wicher is a package used to make working with the WICHE enrollment projections easier in R. We generally use WICHE enrollments to figure out the percentage of students who participate in our assessments.

We also use WICHE to visualize demographics in our Excel dashboards so I'm putting some examples of visualizing in R below. I think this might make it easier to prototype the graphics before trying to work them into the dashboard.

Installing WICHE R Package

You can install wicher from it's github repository like this:

# MAKE SURE YOU HAVE DEVTOOLS INSTALLED
# AND REFERENCED

install.packages("devtools")
library(devtools)

# USE DEVTOOLS TO INSTALL THE WICHE PACKAGE 

install_github('mattjcamp/wicher','mattjcamp')
library(wicher)

# REFERENCE GGPLOT2

install.packages("ggplot2")
library(ggplot2)

You will need to install the packages above if you want to reproduce these examples.

Datasets Included in WICHE R Package

wiche_graduate_projections is the built in dataframe that comes with this package. It looks like this:

library(wicher)
library(tidyverse)

head(wiche_enrollments)

Usually I use this dataframe to JOIN to assessment data aggregated by state and year to get participation rates, but sometimes you may want to just lookup a few datapoints.

Simple Visualizations

So I just started adding visualizations using R since we mostly use Excel to produce graphics. Still, it's nice to visualize data right in R while we are doing analysis. When I'm not too worried about the appearence of the graphics I just use qplot (as in quick plot). This is part of ggplot2 that we added at the top of this doc.

Showing Trends

Here is how to do a quick line plot to show trends. So I pulled the enrollments for White and Hispanic students in California for the years 2000 to 2010. Then I used qplot to make a chart that would show the trend in enrollments for these two groups.

library(ggplot2)

enrollments <- 
  wiche_enrollments %>% 
  filter(year %in% 2000:2010,
         location %in% "ca",
         race %in% c("white", "hispanic"),
         grade == "g") %>% 
  arrange(year, race)

qplot(year, n, data = enrollments,
      color = race,
      geom = c("point", "smooth"),
      main = "California Enrollment",
      xlab = "Academic Year",
      ylab = "Enrollment")

NOTE I found that these plots look terrible unless you set the right figure dimensions in the markdown code. Look at the code file to see how this is done.

Here is the same thing but using the full ggplot function:

ggplot(enrollments, aes(x = as.character(year), y = n, group = race)) +
  geom_smooth(aes(color = race)) +
  geom_point(aes(color = race)) +
  ggtitle("California Enrollment") +
  labs(x = "Academic Year", y = "Enrollment")

NOTE I'm putting this here as a reference for when we want something like this and we want to control the output. ggplot2 has more options that qplot.

Comparing Groups

Here is an example of using column charts to compare groups. We will look again at California at the beginning and the end of the trend examined above.

First, make the chart for 2005:

enrollments <- 
  wiche_enrollments %>% 
  filter(year %in% 2005,
         location %in% "ca",
         race %in% c("white", "asian", "hispanic", "native", "black"),
         grade == "g") %>% 
  arrange(year, race)

ggplot(data = enrollments, aes(race, n)) +
  geom_bar(stat = "identity", position = "dodge") +
  ggtitle("California Enrollment (2005)") +
  labs(x = "Academic Year", y = "Enrollment")

MattjCamp/wicher documentation built on May 8, 2019, 9:52 a.m.