saveyak/arts_education_NOLA.R

## arts_education_NOLA.R
#I wrote this code to produce the graphics used in the public policy report, "Will the Arts Come Marching In? Access to Arts Education in Post-Katrina New Orleans," by Sarah Woodward of Tulane University. Read the report here: https://educationresearchalliancenola.org/files/publications/20200915-Woodward-Will-the-Arts-Come-Marching-in-Access-to-Arts-Education-in-Post-Katrina-New-Orleans.pdf

library(readxl)
library(tidyverse)
library(ggplot2)
library(extrafont)
library(scales)
library(grid)
library(stringr) #lets you use str_wrap()

###Themes and palettes####

era_palette = c("#073b47", "#a8c42e", "#d65C2A", "#3BBAE0", "#7C878E", "#e59918")
blue_gradient = c("#073b47", "#145A6D","#217A93","#2e9ab9","#3bbae0")
#http://www.perbang.dk/rgbgradient/

horizontal_theme = theme_bw(base_family = "Georgia") +
  theme(
    panel.border = element_rect(color = "black", fill=NA),
    axis.title.y = element_text(hjust=0.5, size = 12, margin=margin(l=10, r=10)),
    axis.text.y = element_text(size=12, margin=margin(r=10), color="black", hjust=0),
    axis.text.x = element_text(size=12, margin=margin(t=10, b=10), color="black"),
    axis.title.x = element_blank(),
#    axis.title.x = element_text(hjust=0.5, size = 12, margin=margin(t=10, b=10)),
    legend.title = element_blank(),
    legend.position= "bottom",
    legend.margin=margin(-10, 10, 0, 0),
    legend.text = element_text(size = 12, margin = margin(r = 10)),
    panel.grid.major.x = element_blank() ,
    panel.grid.minor.x = element_blank(),
    panel.grid.major.y = element_line(size=1),
    panel.grid.minor.y = element_blank(),
    plot.margin = margin(30, 10, 10, 10, "pt"),
    strip.text.x = element_text(size = 12))

#Data####

#Note: The dataset include non-publicly-available information on individual schools, and therefore I will not share it online.
data = "arts_ed_data.xlsx"

nola_k8_2016 = read_excel(data, sheet=1)
nola_hs_2016 = read_excel(data, sheet=2)
comparison_k8 = read_excel(data, sheet=3)
comparison_hs = read_excel(data, sheet=4)
talent = read_excel(data, sheet=7)

nola_k8_filtered = nola_k8_2016 %>%  filter(!is.na(sps_lvl)) %>%  filter(sps_lvl != "T")

#Figure 1####

#Figure 1: Fewer K-8 students in New Orleans were enrolled in arts and enrichment courses than in Caddo or East Baton Rouge Parishes.

comparison_k8$district = factor(comparison_k8$district, levels=unique(comparison_k8$district))

ggplot(comparison_k8, aes(x=year, y=pct_artenrich, color=district, group=district)) +
  geom_line(size=1) +
  geom_point(size=3) +
  theme_bw(base_family = "Georgia") +
  horizontal_theme +
  labs(y = "% of K-8 Students\nin Arts or Enrichment Courses") +
  scale_y_continuous(expand = c(0, 0),
                     limits=c(0,100),
                     breaks = seq(0, 100, by=20),
                     label = function(x) paste0(x,"%")) +
  scale_color_manual(values=era_palette)

ggsave("formatted_figs/art_fig1_formatted.pdf", w=6.5, h=3.5, unit="in", device=cairo_pdf)

#Figure 2####

#Figure 2: For high school students, arts enrollment rates were similar across the three parishes.

comparison_hs_nojeff = comparison_hs %>% filter(district != "Jefferson Parish")

comparison_hs_nojeff$district = factor(comparison_hs_nojeff$district, levels=unique(comparison_hs_nojeff$district))

ggplot(comparison_hs_nojeff, aes(x=year, y=pct_arts_course, color=district, group=district)) +
    geom_line(size=1) +
    geom_point(size=3) +
    horizontal_theme +
    labs(y = "% of High School Students\nin Arts Courses",
         title = "") +
    scale_y_continuous(expand = c(0, 0),
                       limits=c(0,100),
                       breaks = seq(0, 100, by=20),
                       label = function(x) paste0(x,"%")) +
    scale_color_manual(values=era_palette)

ggsave("formatted_figs/art_fig2_formatted.pdf", w=6.5, h=3.5, unit="in", device=cairo_pdf)

#Figure 3####

#Figure 3: There is no relationship between school letter grades and enrollment in arts and enrichment.

#This is the average percent enrollment at each school, not a weighted average

k8_sps = nola_k8_filtered %>% group_by(sps_lvl) %>% summarize(mean_enrollment = mean(pct_artenrich))

k8_sps$grade = "K-8"

hs_sps = nola_hs_2016 %>% group_by(sps_lvl) %>% summarize(mean_enrollment = mean(pct_arts_course))

hs_sps$grade = "High School"

both_sps = rbind(k8_sps, hs_sps)

both_sps$grade = factor(both_sps$grade, levels = c("K-8", "High School"))

ggplot(both_sps, aes(x=grade, y=mean_enrollment, fill=sps_lvl))+
  geom_bar(stat="identity", width=0.9, position=position_dodge(width=0.9)) +
  horizontal_theme +
  labs(x = "School Performance Score", y = "Average % Enrollment\nin Arts or EnrichmentCourses", title="") +
  scale_y_continuous(expand = c(0, 0),
                     limits=c(0,100),
                     breaks = seq(0, 100, by=20),
                     labels = function(x) paste0(x, "%")) +
  scale_fill_manual(values = blue_gradient) +
  geom_text(aes(label=paste0(round(mean_enrollment), "%")), size=4.2333, position=position_dodge(width=0.9), family = "Georgia", hjust=0.5, vjust=-0.5) +
  labs(fill="School Performance Score:") + #Changes legend title
  theme(legend.position = "bottom", legend.title = element_text())

ggsave("formatted_figs/art_fig3_formatted.pdf", w=6.5, h=3.5, unit="in", device=cairo_pdf)

#Figure 4####

#Figure 4: Arts enrollment rates vary widely across schools, but standalone charter schools had higher rates than network charters.

#Note: In this case I am *not* filtering out schools with no letter grades

no_OPSB_k8 = nola_k8_2016 %>% filter(nola_type16 != "OPSB direct-run")
no_OPSB_hs = nola_hs_2016 %>% filter(nola_type16 != "OPSB direct-run")

no_OPSB_hs = no_OPSB_hs %>% rename(pct_artenrich = pct_arts_course)

no_OPSB_both = rbind(no_OPSB_k8, no_OPSB_hs)

#Drop Cohen which is incorrectly listed as K-8
no_OPSB_both = no_OPSB_both %>% filter(grades != "DROP THIS")

no_OPSB_both$grades = factor(no_OPSB_both$grades, levels = unique(no_OPSB_both$grades))

no_OPSB_both$nola_type16 = ifelse(no_OPSB_both$nola_type16 == "Standalone charter", "Standalone", "Network")

ggplot(no_OPSB_both, aes(y=pct_artenrich, x=nola_type16)) +
  geom_point(size=2, color="#e59918") +
  scale_color_manual(values=era_palette) +
  theme_bw(base_family = "Georgia") +
  scale_y_continuous(limits=c(0,100),
                     breaks = seq(0, 100, by=20),
                     labels = function(x) paste0(x, "%")) +
  horizontal_theme +
  labs(y = "% of Students in Arts or Enrichment") +
  stat_summary(fun.y=mean, geom="point", shape=18,
               size=4, color="#073b47") +
  stat_summary(aes(label=paste0(round(..y..),"%")), fun.y=mean, geom="text", size=3.52778, hjust=-0.5, family="Georgia") +
  facet_grid(cols = vars(grades))

#If you want the word average: stat_summary(aes(label=paste0("Average: ", round(..y..),"%") %>% str_wrap(width=8)), fun.y=mean, geom="text", size=3.52778, hjust=-0.1, family="Georgia")

#Or alternatively: labels = paste0("Average: ", scales::percent(VARNAME, accuracy = 1)) %>% stringr::str_wrap(width = 8)

#ggsave("formatted_figs/art_fig4_test.pdf", w=6.5, h=4.5, unit="in", device=cairo_pdf)

ggsave("formatted_figs/art_fig4_formatted.pdf", w=6.5, h=4.5, unit="in", device=cairo_pdf)

View(no_OPSB_both)

#Figure 5####
#Figure 5: A-rated schools had a far higher percentage of students identified as artistically talented.

View(talent)

talent$pct_talented = talent$pct_talented*100
talent$grade = factor(talent$grade, levels=c("K-8", "High School"))

ggplot(talent, aes(x=grade, y=pct_talented, fill=SPS))+
  geom_bar(stat="identity", width=0.9, position=position_dodge(width=0.9)) +
  horizontal_theme +
  labs(x = "School Performance Score", y = "% of Students Identified as Talented", title="") +
  scale_y_continuous(expand = c(0, 0),
                     limits=c(0,10),
                     breaks = seq(0, 10, by=2),
                     labels = function(x) paste0(x, "%")) +
  scale_fill_manual(values = blue_gradient) +
  geom_text(aes(label=paste0(pct_talented, "%")), size=3.52778, position=position_dodge(width=0.9), family = "Georgia", hjust=0.5, vjust=-0.5) +
  labs(fill="School Performance Score:") + #Changes legend title
  theme(legend.position = "bottom", legend.title = element_text())

ggsave("formatted_figs/art_fig5_formatted.pdf", w=6.5, h=3.5, unit="in", device=cairo_pdf)
	#I wrote this code to produce the graphics used in the public policy report, "Will the Arts Come Marching In? Access to Arts Education in Post-Katrina New Orleans," by Sarah Woodward of Tulane University. Read the report here: https://educationresearchalliancenola.org/files/publications/20200915-Woodward-Will-the-Arts-Come-Marching-in-Access-to-Arts-Education-in-Post-Katrina-New-Orleans.pdf

	library(readxl)
	library(tidyverse)
	library(ggplot2)
	library(extrafont)
	library(scales)
	library(grid)
	library(stringr) #lets you use str_wrap()

	###Themes and palettes####

	era_palette = c("#073b47", "#a8c42e", "#d65C2A", "#3BBAE0", "#7C878E", "#e59918")
	blue_gradient = c("#073b47", "#145A6D","#217A93","#2e9ab9","#3bbae0")
	#http://www.perbang.dk/rgbgradient/

	horizontal_theme = theme_bw(base_family = "Georgia") +
	theme(
	panel.border = element_rect(color = "black", fill=NA),
	axis.title.y = element_text(hjust=0.5, size = 12, margin=margin(l=10, r=10)),
	axis.text.y = element_text(size=12, margin=margin(r=10), color="black", hjust=0),
	axis.text.x = element_text(size=12, margin=margin(t=10, b=10), color="black"),
	axis.title.x = element_blank(),
	# axis.title.x = element_text(hjust=0.5, size = 12, margin=margin(t=10, b=10)),
	legend.title = element_blank(),
	legend.position= "bottom",
	legend.margin=margin(-10, 10, 0, 0),
	legend.text = element_text(size = 12, margin = margin(r = 10)),
	panel.grid.major.x = element_blank() ,
	panel.grid.minor.x = element_blank(),
	panel.grid.major.y = element_line(size=1),
	panel.grid.minor.y = element_blank(),
	plot.margin = margin(30, 10, 10, 10, "pt"),
	strip.text.x = element_text(size = 12))

	#Data####

	#Note: The dataset include non-publicly-available information on individual schools, and therefore I will not share it online.
	data = "arts_ed_data.xlsx"

	nola_k8_2016 = read_excel(data, sheet=1)
	nola_hs_2016 = read_excel(data, sheet=2)
	comparison_k8 = read_excel(data, sheet=3)
	comparison_hs = read_excel(data, sheet=4)
	talent = read_excel(data, sheet=7)

	nola_k8_filtered = nola_k8_2016 %>% filter(!is.na(sps_lvl)) %>% filter(sps_lvl != "T")

	#Figure 1####

	#Figure 1: Fewer K-8 students in New Orleans were enrolled in arts and enrichment courses than in Caddo or East Baton Rouge Parishes.

	comparison_k8$district = factor(comparison_k8$district, levels=unique(comparison_k8$district))

	ggplot(comparison_k8, aes(x=year, y=pct_artenrich, color=district, group=district)) +
	geom_line(size=1) +
	geom_point(size=3) +
	theme_bw(base_family = "Georgia") +
	horizontal_theme +
	labs(y = "% of K-8 Students\nin Arts or Enrichment Courses") +
	scale_y_continuous(expand = c(0, 0),
	limits=c(0,100),
	breaks = seq(0, 100, by=20),
	label = function(x) paste0(x,"%")) +
	scale_color_manual(values=era_palette)

	ggsave("formatted_figs/art_fig1_formatted.pdf", w=6.5, h=3.5, unit="in", device=cairo_pdf)

	#Figure 2####

	#Figure 2: For high school students, arts enrollment rates were similar across the three parishes.

	comparison_hs_nojeff = comparison_hs %>% filter(district != "Jefferson Parish")

	comparison_hs_nojeff$district = factor(comparison_hs_nojeff$district, levels=unique(comparison_hs_nojeff$district))

	ggplot(comparison_hs_nojeff, aes(x=year, y=pct_arts_course, color=district, group=district)) +
	geom_line(size=1) +
	geom_point(size=3) +
	horizontal_theme +
	labs(y = "% of High School Students\nin Arts Courses",
	title = "") +
	scale_y_continuous(expand = c(0, 0),
	limits=c(0,100),
	breaks = seq(0, 100, by=20),
	label = function(x) paste0(x,"%")) +
	scale_color_manual(values=era_palette)

	ggsave("formatted_figs/art_fig2_formatted.pdf", w=6.5, h=3.5, unit="in", device=cairo_pdf)

	#Figure 3####

	#Figure 3: There is no relationship between school letter grades and enrollment in arts and enrichment.

	#This is the average percent enrollment at each school, not a weighted average

	k8_sps = nola_k8_filtered %>% group_by(sps_lvl) %>% summarize(mean_enrollment = mean(pct_artenrich))

	k8_sps$grade = "K-8"

	hs_sps = nola_hs_2016 %>% group_by(sps_lvl) %>% summarize(mean_enrollment = mean(pct_arts_course))

	hs_sps$grade = "High School"

	both_sps = rbind(k8_sps, hs_sps)

	both_sps$grade = factor(both_sps$grade, levels = c("K-8", "High School"))

	ggplot(both_sps, aes(x=grade, y=mean_enrollment, fill=sps_lvl))+
	geom_bar(stat="identity", width=0.9, position=position_dodge(width=0.9)) +
	horizontal_theme +
	labs(x = "School Performance Score", y = "Average % Enrollment\nin Arts or EnrichmentCourses", title="") +
	scale_y_continuous(expand = c(0, 0),
	limits=c(0,100),
	breaks = seq(0, 100, by=20),
	labels = function(x) paste0(x, "%")) +
	scale_fill_manual(values = blue_gradient) +
	geom_text(aes(label=paste0(round(mean_enrollment), "%")), size=4.2333, position=position_dodge(width=0.9), family = "Georgia", hjust=0.5, vjust=-0.5) +
	labs(fill="School Performance Score:") + #Changes legend title
	theme(legend.position = "bottom", legend.title = element_text())

	ggsave("formatted_figs/art_fig3_formatted.pdf", w=6.5, h=3.5, unit="in", device=cairo_pdf)

	#Figure 4####

	#Figure 4: Arts enrollment rates vary widely across schools, but standalone charter schools had higher rates than network charters.

	#Note: In this case I am not filtering out schools with no letter grades

	no_OPSB_k8 = nola_k8_2016 %>% filter(nola_type16 != "OPSB direct-run")
	no_OPSB_hs = nola_hs_2016 %>% filter(nola_type16 != "OPSB direct-run")

	no_OPSB_hs = no_OPSB_hs %>% rename(pct_artenrich = pct_arts_course)

	no_OPSB_both = rbind(no_OPSB_k8, no_OPSB_hs)

	#Drop Cohen which is incorrectly listed as K-8
	no_OPSB_both = no_OPSB_both %>% filter(grades != "DROP THIS")

	no_OPSB_both$grades = factor(no_OPSB_both$grades, levels = unique(no_OPSB_both$grades))

	no_OPSB_both$nola_type16 = ifelse(no_OPSB_both$nola_type16 == "Standalone charter", "Standalone", "Network")

	ggplot(no_OPSB_both, aes(y=pct_artenrich, x=nola_type16)) +
	geom_point(size=2, color="#e59918") +
	scale_color_manual(values=era_palette) +
	theme_bw(base_family = "Georgia") +
	scale_y_continuous(limits=c(0,100),
	breaks = seq(0, 100, by=20),
	labels = function(x) paste0(x, "%")) +
	horizontal_theme +
	labs(y = "% of Students in Arts or Enrichment") +
	stat_summary(fun.y=mean, geom="point", shape=18,
	size=4, color="#073b47") +
	stat_summary(aes(label=paste0(round(..y..),"%")), fun.y=mean, geom="text", size=3.52778, hjust=-0.5, family="Georgia") +
	facet_grid(cols = vars(grades))

	#If you want the word average: stat_summary(aes(label=paste0("Average: ", round(..y..),"%") %>% str_wrap(width=8)), fun.y=mean, geom="text", size=3.52778, hjust=-0.1, family="Georgia")

	#Or alternatively: labels = paste0("Average: ", scales::percent(VARNAME, accuracy = 1)) %>% stringr::str_wrap(width = 8)

	#ggsave("formatted_figs/art_fig4_test.pdf", w=6.5, h=4.5, unit="in", device=cairo_pdf)

	ggsave("formatted_figs/art_fig4_formatted.pdf", w=6.5, h=4.5, unit="in", device=cairo_pdf)

	View(no_OPSB_both)

	#Figure 5####
	#Figure 5: A-rated schools had a far higher percentage of students identified as artistically talented.

	View(talent)

	talent$pct_talented = talent$pct_talented*100
	talent$grade = factor(talent$grade, levels=c("K-8", "High School"))

	ggplot(talent, aes(x=grade, y=pct_talented, fill=SPS))+
	geom_bar(stat="identity", width=0.9, position=position_dodge(width=0.9)) +
	horizontal_theme +
	labs(x = "School Performance Score", y = "% of Students Identified as Talented", title="") +
	scale_y_continuous(expand = c(0, 0),
	limits=c(0,10),
	breaks = seq(0, 10, by=2),
	labels = function(x) paste0(x, "%")) +
	scale_fill_manual(values = blue_gradient) +
	geom_text(aes(label=paste0(pct_talented, "%")), size=3.52778, position=position_dodge(width=0.9), family = "Georgia", hjust=0.5, vjust=-0.5) +
	labs(fill="School Performance Score:") + #Changes legend title
	theme(legend.position = "bottom", legend.title = element_text())

	ggsave("formatted_figs/art_fig5_formatted.pdf", w=6.5, h=3.5, unit="in", device=cairo_pdf)