DAT-4313 - Lab 7.1

School Data

Author

Luke Stucky

library(tidyverse)
library(ggplot2)
library(ggthemes)
library(socviz)
library(maps)
library(mapproj)
library(questionr)
library(viridis)
library(leaflet)
library(tidycensus)
library(knitr)
library(dplyr)
library(readxl)

GET KANSAS POPULATION DATA BY COUNTY DIRECTLY FROM THE US CENSUS VIA API

census_api_key("4e54c9226f6c18a4ef1873e3fa856936b9fa1775")  #use this one when rendering

ks_pop <- 
  get_acs(geography = "county",
          variables = "B01003_001",
          state = "KS",
          geometry = TRUE)

head(as_tibble(ks_pop))

# A tibble: 6 × 6
  GEOID NAME                   variable estimate   moe                  geometry
  <chr> <chr>                  <chr>       <dbl> <dbl>        <MULTIPOLYGON [°]>
1 20175 Seward County, Kansas  B01003_…    21942    NA (((-101.0681 37.38778, -…
2 20027 Clay County, Kansas    B01003_…     8088    NA (((-97.3707 39.21899, -9…
3 20171 Scott County, Kansas   B01003_…     5128    NA (((-101.1284 38.7006, -1…
4 20047 Edwards County, Kansas B01003_…     2849    NA (((-99.56988 37.91302, -…
5 20147 Phillips County, Kans… B01003_…     4960    NA (((-99.62821 39.72017, -…
6 20149 Pottawatomie County, … B01003_…    25482    NA (((-96.72774 39.42685, -…

CHOLOPLETH FOR KANSAS COUNTIES

MapPalette <- colorQuantile(palette = "viridis", domain = ks_pop$estimate, n = 20)

library(sf)
ks_pop %>% 
  st_transform(crs = "+proj=longlat +datum=WGS84") %>% 
  leaflet(width = "100%", height = 500) %>% 
  addProviderTiles(provider = "Esri.WorldPhysical") %>% 
  addPolygons(popup = ~NAME,
              stroke = FALSE,
              smoothFactor = 0,
              fillOpacity = 0.7,
              color = ~ MapPalette(estimate)) %>% 
  addLegend("bottomright", 
            pal = MapPalette,
            values = ~ estimate,
            title = "Population Percentiles",
            opacity = 1)

Kansas is a diverse state in terms of population. Western Kansas is known for having very few people, while Sedgwick County and Johnson County hold a high percentage of the population. The two major Universities in Kansas, Kansas State University and Kansas University reside in Riley County and Johnson County.

SCHOOL LOCATIONS IN KANSAS

Public_Schools <- read.csv("Public_Schools.csv")

Private_Schools <- read.csv("Private_Schools.csv")

public_schools_ks <- Public_Schools %>% 
  filter(STATE == "KS")

private_schools_ks <- Private_Schools %>% 
  filter(STATE == "KS")

# Public Schools
public_schools_ks %>% leaflet(width = "100%") %>% 
  addTiles() %>% 
  setView(-98.0, 39.0, zoom = 6) %>% 
  addMarkers(lat = ~LATITUDE, 
             lng = ~LONGITUDE, 
             popup = ~NAME)

# Private Schools
private_schools_ks %>% leaflet(width = "100%") %>% 
  addTiles() %>% 
  setView(-98.0, 39.0, zoom = 6) %>% 
  addMarkers(lat = ~LATITUDE, 
             lng = ~LONGITUDE, 
             popup = ~NAME)

There are more public schools than private schools, but this is because public schools are split up into multiple schools depending on grade level. Both public and private hold a similar trend as they both seem to cluster a lot in Sedgwick County and Johnson County. As I mentioned previously, these counties hold the highest densities.

Aquired Data from Kaggle (US Schools Dataset - LARXEL).

SCHOOL LOCATIONS OVERLAYED ONTO THE CHOROPLETH

# Used ChatGPT
library(leaflet)
library(sf)

public_schools_ks$type <- "Public"
private_schools_ks$type <- "Private"

public_schools_ks <- public_schools_ks[, c("NAME", "LONGITUDE", "LATITUDE", "ENROLLMENT", "type")]
private_schools_ks <- private_schools_ks[, c("NAME", "LONGITUDE", "LATITUDE", "ENROLLMENT", "type")]

combined_schools <- bind_rows(public_schools_ks, private_schools_ks)

sorted_schools <- combined_schools[order(-combined_schools$ENROLLMENT), ]

class_sizes <- c(16, 16, 32, 48, 48, 48, Inf)

class_names <- c("7", "6", "5", "4", "3", "2", "1")

sorted_schools$class <- cut(x = rank(-sorted_schools$ENROLLMENT, ties.method = "first"),
                            breaks = cumsum(c(0, class_sizes)),
                            labels = class_names,
                            include.lowest = TRUE)

I read online that schools were split up into the 7 classification based on which schools had the highest number of students. Using the enrollment variable, I was able to subset the schools into the 7 classifications. While it is not perfect, it does the job for collecting some valuable insights.

# Used ChatGPT
MapPalette <- colorQuantile(palette = "viridis", domain = ks_pop$estimate, n = 20)


ks_pop %>% 
  st_transform(crs = "+proj=longlat +datum=WGS84") %>% 
  leaflet(width = "100%", height = 500) %>% 
  addProviderTiles(provider = "Esri.WorldPhysical") %>% 
  addPolygons(popup = ~NAME,
              stroke = FALSE,
              smoothFactor = 0,
              fillOpacity = 0.7,
              color = ~ MapPalette(estimate)) %>% 
  addLegend("bottomright", 
            pal = MapPalette,
            values = ~ estimate,
            title = "Population Percentiles",
            opacity = 1) %>% 
  addCircleMarkers(data = sorted_schools,
                   lng = ~LONGITUDE,
                   lat = ~LATITUDE,
                   radius = ~class,
                   color = ~type,
                   fillColor = ~colorFactor(palette = c("red", "blue"), domain = c("Public", "Private"))(type),
                   fillOpacity = 0.7,
                   popup = ~NAME) %>%
  addLegend("bottomleft",
            pal = colorFactor(palette = c("red", "blue"), domain = c("Public", "Private")),
            values = sorted_schools$type,
            title = "School Type",
            opacity = 1)

I decided to separate the type of school (private or public) by the color. Red means that the school is private while blue means that it is public. I also made the size of the circle the classification of the school (1A-7A). The bigger the circle, the higher the classification, and the larger the student body. By doing this, I was able to create some insights. The biggest schools are in the counties with the highest populations. While there are not many big private schools, there seems to be a great deal of them clustered up in the high population counties.