This study establishes a reproducible geospatial screening framework to assess child environmental health vulnerability across Cook County, Chicago, Illinois. By synthesizing tract-level demographic indicators from the American Community Survey (ACS) with PM₂.₅ exposure data from OpenAQ, the analysis captures the interactions among biological (host), environmental, and social domains that influence child health.
A composite vulnerability index was developed to identify areas where pediatric sensitivity and environmental burden converge. Inverse Distance Weighting (IDW) was applied to address spatial gaps in monitoring coverage and to generate a continuous PM₂.₅ surface, enabling a fully spatially referenced assessment of exposure.
The results indicate pronounced spatial disparities across Chicago’s South and West Sides, where elevated air pollution coincides with higher poverty rates and greater concentrations of young children. The resulting maps and interactive web application offer a scalable, data-driven framework to support targeted public health interventions and promote equity-focused decision-making in child environmental health.
The primary objective of this study is to develop a multi-domain assessment of child environmental health vulnerability by identifying census tracts where pediatric sensitivity coincides with environmental exposure.
The analysis specifically evaluates three interrelated dimensions:
Biological (host) vulnerability, defined as the concentration of children under age five.
Social domain vulnerability, measured by the proportion of the population living below the poverty line.
Environmental exposure, represented by ambient PM₂.₅ concentration.
Integration of these domains into a composite vulnerability index enables quantification of spatial disparities and identification of priority areas where environmental and social risks converge, thereby supporting targeted interventions and place-based public health strategies.
This project integrates multiple public datasets:
a) American Community Survey (ACS), 2023
Total population (B01003_001)
Male under age 5 (B01001_003)
Female under age 5 (B01001_027)
Population below poverty (B17001_002)
b) OpenAQ API
c) Tiger/Line /Tigris boundary data
Chicago City Boundary
Census tract geometries for Cook County, Illinois
library(tidycensus)
library(tidyverse)
library(sf)
library(tmap)
library(janitor)
library(viridis)
library(httr)
library(jsonlite)
library(dplyr)
library(leaflet)
library(scales)
library(tigris)
library(leaflet)
library(leaflet.extras)
library(sf)
library(dplyr)
library(scales)
library(htmltools)
library(knitr)
library(kableExtra)
library(gt)
library (DT)
options(tigris_use_cache = TRUE)
tmap_mode("plot")American Community Survey (ACS) tract-level data were obtained for Cook County, Illinois. The variables total population, children under five, and poverty were utilized to calculate percentage-based indicators.
CENSUS_API_KEY <- "312248c846fdb30816100611e375688ee60be39f"
census_api_key(CENSUS_API_KEY, install = FALSE, overwrite = TRUE)
state_name <- "IL"
county_name <- "Cook"
year_use <- 2023
acs_vars <- c(
total_pop = "B01003_001",
male_under5 = "B01001_003",
female_under5 = "B01001_027",
below_poverty = "B17001_002"
)
acs_raw <- get_acs(
geography = "tract",
variables = acs_vars,
state = state_name,
county = county_name,
year = year_use,
geometry = TRUE,
output = "wide"
)
acs_data <- acs_raw %>%
transmute(
geoid = GEOID,
name = NAME,
total_pop = total_popE,
male_under5 = male_under5E,
female_under5 = female_under5E,
below_poverty = below_povertyE,
geometry = geometry
) %>%
mutate(
children_under5 = male_under5 + female_under5,
pct_children_under5 = if_else(
total_pop > 0,
100 * children_under5 / total_pop,
NA_real_
),
pct_below_poverty = if_else(
total_pop > 0,
100 * below_poverty / total_pop,
NA_real_
)
) %>%
st_transform(crs = 4326)The analysis was limited to census tracts located within the boundaries of Chicago.
chicago_boundary <- places(state = "IL", cb = TRUE) %>%
filter(NAME == "Chicago") %>%
st_transform(crs = 4326)
acs_chicago <- acs_data %>%
st_filter(chicago_boundary, .predicate = st_within)PM2.5 monitoring locations were obtained via the OpenAQ API. These locations were restricted to the Chicago city boundary, and only valid PM2.5 observations were included in the analysis.
# OpenAQ API key
OPENAQ_KEY <- "d9e539fa59e9ad36ba0c97863b47b6f9e5ee0ce7abe3507fed8d796e05ee59ad"
loc_res <- GET(
"https://api.openaq.org/v3/locations",
query = list(
bbox = "-88.0,41.6,-87.5,42.1",
parameters_id = 2,
limit = 100
),
add_headers("X-API-Key" = OPENAQ_KEY)
)
locations <- fromJSON(
content(loc_res, "text", encoding = "UTF-8"),
flatten = TRUE
)$results
locations_sf <- locations %>%
filter(
!is.na(coordinates.latitude),
!is.na(coordinates.longitude)
) %>%
st_as_sf(
coords = c("coordinates.longitude", "coordinates.latitude"),
crs = 4326
)
chicago_monitors <- locations_sf %>%
st_filter(chicago_boundary, .predicate = st_within)PM2.5 sensor data were collected from all selected monitoring sites in Chicago. Data cleaning procedures were applied to retain only plausible PM2.5 measurements.
get_sensors <- function(loc_id) {
res <- GET(
paste0("https://api.openaq.org/v3/locations/", loc_id, "/sensors"),
add_headers("X-API-Key" = OPENAQ_KEY)
)
if (status_code(res) != 200) return(NULL)
d <- fromJSON(
content(res, "text", encoding = "UTF-8"),
flatten = TRUE
)$results
if (is.null(d) || length(d) == 0) return(NULL)
d <- as.data.frame(d)
d$location_id <- loc_id
d
}
sensor_list <- lapply(chicago_monitors$id, function(id) {
Sys.sleep(0.2)
get_sensors(id)
})
sensors_df <- bind_rows(sensor_list)
pm25_df <- sensors_df %>%
filter(
parameter.name == "pm25",
!is.na(latest.value),
latest.value > 0,
latest.value < 200
) %>%
select(
location_id,
pm25_mean = latest.value,
pm25_lat = latest.coordinates.latitude,
pm25_lon = latest.coordinates.longitude
) %>%
group_by(location_id) %>%
summarise(
pm25_mean = mean(pm25_mean, na.rm = TRUE),
pm25_lat = first(pm25_lat),
pm25_lon = first(pm25_lon),
.groups = "drop"
) %>%
filter(!is.na(pm25_lat), !is.na(pm25_lon)) %>%
left_join(
chicago_monitors %>%
st_drop_geometry() %>%
select(id, name),
by = c("location_id" = "id")
)
aq_points <- pm25_df %>%
st_as_sf(coords = c("pm25_lon", "pm25_lat"), crs = 4326)A diagnostic map was generated to assess the spatial distribution of monitor locations and PM2.5 measurements in Chicago prior to interpolation.
pal_diag <- colorNumeric(
palette = c("#00e400", "#ffff00", "#ff7e00", "#ff0000"),
domain = pm25_df$pm25_mean
)
diagnostic_map <- leaflet() %>%
addProviderTiles(providers$CartoDB.Positron) %>%
addPolygons(
data = chicago_boundary,
fillColor = "transparent",
color = "#1a4f72",
weight = 2.5,
dashArray = "5,5",
group = "Chicago Boundary"
) %>%
addCircleMarkers(
data = pm25_df,
lng = ~pm25_lon,
lat = ~pm25_lat,
radius = ~rescale(pm25_mean, to = c(6, 18)),
fillColor = ~pal_diag(pm25_mean),
fillOpacity = 0.85,
color = "white",
weight = 1.5,
group = "PM2.5 Monitors",
popup = ~paste0(
"<div style='font-family:sans-serif; min-width:180px'>",
"<b style='color:#1a4f72'>", name, "</b><hr>",
"📍 ID: <b>", location_id, "</b><br>",
"🌫️ PM2.5: <b>", round(pm25_mean, 1), " µg/m³</b><br>",
ifelse(pm25_mean <= 12, "🟢 Good",
ifelse(pm25_mean <= 35, "🟡 Moderate",
ifelse(pm25_mean <= 55, "🟠 Sensitive Groups",
"🔴 Unhealthy"))),
"</div>"
),
label = ~paste0(name, ": ", round(pm25_mean, 1), " µg/m³")
) %>%
addLegend(
position = "bottomright",
pal = pal_diag,
values = pm25_df$pm25_mean,
title = "PM2.5 (µg/m³)",
opacity = 0.9
) %>%
addControl(
html = "
<div style='background:white;padding:10px;border-radius:6px;
font-size:12px;line-height:1.8'>
<b>EPA Air Quality Index</b><br>
🟢 Good: 0–12 µg/m³<br>
🟡 Moderate: 12–35 µg/m³<br>
🟠 Sensitive: 35–55 µg/m³<br>
🔴 Unhealthy: 55+ µg/m³
</div>",
position = "topleft"
) %>%
addLayersControl(
overlayGroups = c("Chicago Boundary", "PM2.5 Monitors"),
options = layersControlOptions(collapsed = FALSE)
) %>%
addScaleBar(position = "bottomleft") %>%
setView(lng = -87.70, lat = 41.85, zoom = 11)
diagnostic_mapPM2.5 monitor values were joined to census tracts. Because many tracts lacked a monitor, inverse distance weighting (IDW) was used to estimate PM2.5 concentrations for unmonitored tracts.
aq_in_tracts <- st_join(aq_points, acs_chicago, join = st_within)
tract_pm25 <- aq_in_tracts %>%
st_drop_geometry() %>%
group_by(geoid) %>%
summarise(avg_pm25 = mean(pm25_mean, na.rm = TRUE), .groups = "drop")
acs_chicago <- acs_chicago %>%
left_join(tract_pm25, by = "geoid")
tract_centroids <- acs_chicago %>%
st_centroid() %>%
select(geoid)
idw_pm25 <- function(tract_geom, monitors, power = 2) {
dists <- as.numeric(st_distance(tract_geom, monitors))
dists[dists == 0] <- 1
weights <- 1 / (dists ^ power)
weighted.mean(monitors$pm25_mean, weights, na.rm = TRUE)
}
interpolated_values <- sapply(
seq_len(nrow(tract_centroids)),
function(i) idw_pm25(tract_centroids[i, ], aq_points)
)
acs_chicago$pm25_interpolated <- interpolated_values
acs_chicago <- acs_chicago %>%
mutate(
pm25_final = case_when(
!is.na(avg_pm25) ~ avg_pm25,
!is.na(pm25_interpolated) ~ pm25_interpolated,
TRUE ~ NA_real_
),
pm25_source = case_when(
!is.na(avg_pm25) ~ "Monitor",
!is.na(pm25_interpolated) ~ "Interpolated (IDW)",
TRUE ~ "No Data"
)
)A final tract-level environmental vulnerability index was created by averaging standardized values for:
percentage of children under age five,
percentage below poverty, and
PM2.5 concentration.
acs_chicago <- acs_chicago %>%
mutate(
z_children = as.numeric(scale(pct_children_under5)),
z_poverty = as.numeric(scale(pct_below_poverty)),
z_pm25 = as.numeric(scale(pm25_final)),
env_vulnerability = (z_children + z_poverty + z_pm25) / 3,
risk_tier = cut(
env_vulnerability,
breaks = quantile(
env_vulnerability,
probs = c(0, .25, .5, .75, 1),
na.rm = TRUE
),
labels = c("Low", "Moderate", "High", "Very High"),
include.lowest = TRUE
)
)
acs_chicago %>%
st_drop_geometry() %>%
select(
geoid, name, total_pop, children_under5,
pct_children_under5, below_poverty,
pct_below_poverty, pm25_final,
env_vulnerability, risk_tier
) %>%
mutate(
pct_children_under5 = round(pct_children_under5, 1),
pct_below_poverty = round(pct_below_poverty, 1),
pm25_final = round(pm25_final, 1),
env_vulnerability = round(env_vulnerability, 2)
) %>%
datatable(
colnames = c(
"GEOID", "Census Tract", "Total Pop",
"Children <5", "% Children <5",
"Below Poverty", "% Below Poverty",
"PM2.5 (µg/m³)", "Vulnerability Index", "Risk Tier"
),
rownames = FALSE,
filter = "top",
extensions = "Scroller",
options = list(
pageLength = 5,
lengthMenu = c(5, 10, 25, 50),
scrollX = TRUE,
scrollY = "280px",
scroller = TRUE,
deferRender = TRUE,
dom = "Bfrtip",
buttons = list(
list(extend = "csv", text = "⬇ CSV"),
list(extend = "excel", text = "⬇ Excel")
),
columnDefs = list(
list(width = "180px", targets = 1),
list(width = "80px", targets = c(2, 3, 4, 5, 6, 7, 8)),
list(className = "dt-center", targets = c(2, 3, 4, 5, 6, 7, 8, 9))
),
initComplete = JS("
function(settings, json) {
$(this.api().table().header()).css({
'background-color': '#1a4f72',
'color': 'white',
'font-weight': 'bold',
'font-size': '13px'
});
$(this.api().table().node()).css({
'font-size': '13px',
'font-family': 'Inter, sans-serif'
});
}
")
),
callback = JS("
table.on('draw', function() {
table.rows().every(function(i) {
var tier = this.data()[9];
var colors = {
'Very High': '#f8d7da',
'High': '#ffeeba',
'Moderate': '#d1ecf1',
'Low': '#d4edda'
};
if (colors[tier]) {
$(this.node()).css('background-color', colors[tier]);
}
});
});
")
)tmap_options(component.autoscale = TRUE)
map1 <- tm_shape(acs_chicago) +
tm_polygons(
"pct_children_under5",
palette = "YlGnBu",
style = "quantile",
title = "% Children Under 5"
) +
tm_borders(col = "grey60", lwd = 0.15) +
tm_layout(
title = "Children Under 5 — Chicago Census Tracts",
title.size = 1.4,
title.fontface = "bold",
title.position = c("center", "top"),
frame = TRUE,
bg.color = "white",
legend.outside = TRUE,
legend.outside.position = "right",
inner.margins = c(0.09, 0.02, 0.14, 0.02)
) +
tm_scale_bar(
position = c(0.12, 0.085),
text.size = 0.6,
lwd = 0.6
) +
tm_compass(
position = c(0.88, 0.88),
type = "arrow",
size = 1.8
)
map1
map2 <- tm_shape(acs_chicago) +
tm_polygons(
"pct_below_poverty",
palette = "BuGn",
style = "quantile",
title = "% Below Poverty"
) +
tm_borders(col = "grey60", lwd = 0.2) +
tm_layout(
title = "Poverty Rate - Chicago Census Tracts",
title.size = 2,
title.fontface = "bold",
title.position = c("center", "top"),
frame = TRUE,
bg.color = "white",
legend.outside = TRUE,
legend.outside.position = "right",
inner.margins = c(0.09, 0.02, 0.14, 0.02)
) +
tm_scale_bar(
position = c(0.12, 0.08),
text.size = 0.6,
lwd = 0.6
) +
tm_compass(
position = c(0.88, 0.88),
type = "arrow",
size = 1.8
)
map2
map3 <- tm_shape(acs_chicago) +
tm_polygons(
col = "#f4a582",
border.col = "grey60",
lwd = 0.2
) +
tm_shape(aq_points) +
tm_symbols(
shape = 21,
size = 0.35,
col = "red",
fill = "red",
border.col = "red"
) +
tm_add_legend(
type = "fill",
labels = c("Interpolated (IDW)", "Monitor Stations"),
col = c("#f4a582", "red"),
title = "PM2.5 Source"
) +
tm_layout(
title = "PM2.5 Data Coverage — Chicago Census Tracts",
title.size = 2,
title.fontface = "bold",
title.position = c("center", "top"),
frame = TRUE,
bg.color = "white",
legend.outside = TRUE,
legend.outside.position = "right",
inner.margins = c(0.09, 0.02, 0.14, 0.02)
) +
tm_scale_bar(
position = c(0.12, 0.08),
text.size = 0.6,
lwd = 0.6
) +
tm_compass(
position = c(0.88, 0.88),
type = "arrow",
size = 1.8
)
map3
map4 <- tm_shape(acs_chicago) +
tm_polygons(
"pm25_final",
palette = c("#00e400", "#ffff00", "#ff7e00", "#ff0000", "#8f3f97"),
title = "PM2.5 (µg/m³)",
style = "quantile",
colorNA = "white", # hide NA but keep extent
textNA = "" # no "Missing" label
) +
tm_layout(
title = "PM2.5 Concentration Surface (IDW) — Chicago Census Tracts",
title.size = 1.4,
title.fontface = "bold",
title.position = c("center", "top"),
frame = TRUE,
bg.color = "white",
legend.outside = TRUE,
legend.outside.position = "right",
inner.margins = c(0.02, 0.01, 0.08, 0.01)
) +
tm_scale_bar(
position = c(0.12, 0.08),
text.size = 0.6,
lwd = 0.6
) +
tm_compass(
position = c(0.88, 0.88),
type = "arrow",
size = 1.4
)
map4
map5 <- tm_shape(acs_chicago) +
tm_polygons(
"env_vulnerability",
palette = "-RdYlBu",
title = "Vulnerability Index",
midpoint = 0,
style = "quantile",
colorNA = "white",
textNA = ""
) +
tm_layout(
title = "Child Environmental Vulnerability — Chicago Census Tracts",
title.size = 1.4,
title.fontface = "bold",
title.position = c("center", "top"),
frame = TRUE,
bg.color = "white",
legend.outside = TRUE,
legend.outside.position = "right",
inner.margins = c(0.02, 0.01, 0.08, 0.01)
) +
tm_scale_bar(
position = c(0.12, 0.08),
text.size = 0.6,
lwd = 0.6
) +
tm_compass(
position = c(0.88, 0.88),
type = "arrow",
size = 1.6
)
map5
summary_table <- acs_chicago %>%
st_drop_geometry() %>%
summarise(
`Total Tracts` = n(),
`Avg % Children Under 5` = round(mean(pct_children_under5, na.rm=TRUE), 1),
`Avg % Below Poverty` = round(mean(pct_below_poverty, na.rm=TRUE), 1),
`Avg PM2.5 (µg/m³)` = round(mean(pm25_final, na.rm=TRUE), 1),
`Very High Risk Tracts` = sum(risk_tier == "Very High", na.rm=TRUE),
`High Risk Tracts` = sum(risk_tier == "High", na.rm=TRUE),
`Tracts with Real Monitor` = sum(pm25_source == "Monitor", na.rm=TRUE),
`Tracts IDW Interpolated` = sum(pm25_source == "Interpolated (IDW)", na.rm=TRUE)
) %>%
pivot_longer(
cols = everything(),
names_to = "Indicator",
values_to = "Value"
) %>%
mutate(
Category = c(
"Study Area",
"Demographics",
"Demographics",
"Air Quality",
"Vulnerability",
"Vulnerability",
"Air Quality",
"Air Quality"
),
Unit = c(
"tracts",
"%",
"%",
"µg/m³",
"tracts",
"tracts",
"tracts",
"tracts"
)
)
# Render
summary_table %>%
gt() %>%
cols_label(
Category = "Category",
Indicator = "Indicator",
Value = "Value",
Unit = "Unit"
) %>%
tab_header(
title = "Summary Statistics — Child Environmental Health Indicators",
subtitle = "Chicago Census Tracts · ACS 2023 & OpenAQ"
) %>%
tab_style(
style = list(
cell_fill(color = "#1a4f72"),
cell_text(color = "white", weight = "bold")
),
locations = cells_title(groups = "title")
) %>%
tab_style(
style = list(
cell_fill(color = "#2e86ab"),
cell_text(color = "white", weight = "bold")
),
locations = cells_column_labels()
) %>%
tab_row_group(label = " Vulnerability", rows = Category == "Vulnerability") %>%
tab_row_group(label = " Air Quality", rows = Category == "Air Quality") %>%
tab_row_group(label = "Demographics", rows = Category == "Demographics") %>%
tab_row_group(label = " Study Area", rows = Category == "Study Area") %>%
cols_hide(columns = Category) %>%
tab_style(
style = cell_fill(color = "#f8d7da"),
locations = cells_body(rows = Category == "Vulnerability")
) %>%
tab_style(
style = cell_fill(color = "#eaf4fb"),
locations = cells_body(rows = Category == "Air Quality")
) %>%
tab_style(
style = cell_fill(color = "#f0f8f0"),
locations = cells_body(rows = Category == "Demographics")
) %>%
cols_align(align = "center", columns = c(Value, Unit)) %>%
cols_align(align = "left", columns = Indicator) %>%
tab_footnote(
footnote = "IDW = Inverse Distance Weighting interpolation for unmonitored tracts.",
locations = cells_column_labels(columns = Indicator)
) %>%
opt_table_font(font = "Inter") %>%
opt_row_striping(row_striping = FALSE)| Summary Statistics — Child Environmental Health Indicators | ||
|---|---|---|
| Chicago Census Tracts · ACS 2023 & OpenAQ | ||
| Indicator1 | Value | Unit |
| Study Area | ||
| Total Tracts | 789.0 | tracts |
| Demographics | ||
| Avg % Children Under 5 | 5.6 | % |
| Avg % Below Poverty | 18.0 | % |
| Air Quality | ||
| Avg PM2.5 (µg/m³) | 23.7 | µg/m³ |
| Tracts with Real Monitor | 60.0 | tracts |
| Tracts IDW Interpolated | 729.0 | tracts |
| Vulnerability | ||
| Very High Risk Tracts | 197.0 | tracts |
| High Risk Tracts | 196.0 | tracts |
| 1 IDW = Inverse Distance Weighting interpolation for unmonitored tracts. | ||
# WGS84 for leaflet
acs_web <- st_transform(acs_chicago, 4326)
aq_points_web <- st_transform(aq_points, 4326)
boundary_web <- st_transform(chicago_boundary, 4326)
# Safe label field for monitor stations
aq_points_web$label_name <- if ("name" %in% names(aq_points_web)) {
aq_points_web$name
} else {
rep("Monitor Station", nrow(aq_points_web))
}
# Color palettes
pal_children <- colorQuantile(
palette = "YlGnBu",
domain = acs_web$pct_children_under5,
n = 5,
na.color = "transparent"
)
pal_poverty <- colorQuantile(
palette = "BuGn",
domain = acs_web$pct_below_poverty,
n = 5,
na.color = "transparent"
)
pal_pm25_surface <- colorQuantile(
palette = c("#00e400", "#ffff00", "#ff7e00", "#ff0000", "#8f3f97"),
domain = acs_web$pm25_final,
n = 5,
na.color = "white"
)
pal_vulnerability <- colorQuantile(
palette = rev(RColorBrewer::brewer.pal(11, "RdYlBu")),
domain = acs_web$env_vulnerability,
n = 5,
na.color = "white"
)
pal_monitor <- colorNumeric(
palette = c("#00e400", "#ffff00", "#ff7e00", "#ff0000", "#8f3f97"),
domain = aq_points_web$pm25_mean,
na.color = "transparent"
)
# Popups
popup_children <- paste0(
"<div style='font-family:sans-serif; min-width:220px;'>",
"<b>Tract:</b> ", htmlEscape(acs_web$name), "<br>",
"<b>GEOID:</b> ", acs_web$geoid, "<br>",
"<b>Total Population:</b> ", scales::comma(acs_web$total_pop), "<br>",
"<b>Children Under 5:</b> ", scales::comma(acs_web$children_under5), "<br>",
"<b>% Children Under 5:</b> ", round(acs_web$pct_children_under5, 2), "%",
"</div>"
)
popup_poverty <- paste0(
"<div style='font-family:sans-serif; min-width:220px;'>",
"<b>Tract:</b> ", htmlEscape(acs_web$name), "<br>",
"<b>GEOID:</b> ", acs_web$geoid, "<br>",
"<b>Population Below Poverty:</b> ", scales::comma(acs_web$below_poverty), "<br>",
"<b>% Below Poverty:</b> ", round(acs_web$pct_below_poverty, 2), "%",
"</div>"
)
popup_pm25_source <- paste0(
"<div style='font-family:sans-serif; min-width:220px;'>",
"<b>Tract:</b> ", htmlEscape(acs_web$name), "<br>",
"<b>GEOID:</b> ", acs_web$geoid, "<br>",
"<b>PM2.5 Source:</b> ", acs_web$pm25_source, "<br>",
"<b>Final PM2.5:</b> ", round(acs_web$pm25_final, 2), " µg/m³",
"</div>"
)
popup_pm25_surface <- paste0(
"<div style='font-family:sans-serif; min-width:220px;'>",
"<b>Tract:</b> ", htmlEscape(acs_web$name), "<br>",
"<b>GEOID:</b> ", acs_web$geoid, "<br>",
"<b>PM2.5 Concentration:</b> ", round(acs_web$pm25_final, 2), " µg/m³",
"</div>"
)
popup_vulnerability <- paste0(
"<div style='font-family:sans-serif; min-width:220px;'>",
"<b>Tract:</b> ", htmlEscape(acs_web$name), "<br>",
"<b>GEOID:</b> ", acs_web$geoid, "<br>",
"<b>% Children Under 5:</b> ", round(acs_web$pct_children_under5, 2), "%<br>",
"<b>% Below Poverty:</b> ", round(acs_web$pct_below_poverty, 2), "%<br>",
"<b>PM2.5:</b> ", round(acs_web$pm25_final, 2), " µg/m³<br>",
"<b>Vulnerability Index:</b> ", round(acs_web$env_vulnerability, 2),
"</div>"
)
popup_monitors <- paste0(
"<div style='font-family:sans-serif; min-width:200px;'>",
"<b style='color:#1a4f72; font-size:14px;'>",
htmlEscape(aq_points_web$label_name), "</b><br>",
"<hr style='margin:5px 0;'>",
if ("location_id" %in% names(aq_points_web)) {
paste0("📍 <b>ID:</b> ", aq_points_web$location_id, "<br>")
} else "",
if ("pm25_mean" %in% names(aq_points_web)) {
paste0(
"🌫️ <b>PM2.5:</b> <span style='color:",
pal_monitor(aq_points_web$pm25_mean),
"; font-weight:bold;'>",
round(aq_points_web$pm25_mean, 2), " µg/m³</span><br>"
)
} else "",
ifelse(
aq_points_web$pm25_mean <= 12, "🟢 Good",
ifelse(
aq_points_web$pm25_mean <= 35, "🟡 Moderate",
ifelse(
aq_points_web$pm25_mean <= 55, "🟠 Unhealthy for Sensitive Groups",
"🔴 Unhealthy"
)
)
),
"</div>"
)
# Final interactive map
leaflet(options = leafletOptions(zoomControl = TRUE, minZoom = 9)) %>%
# Base maps
addProviderTiles(providers$CartoDB.Positron,
group = "CartoDB Positron") %>%
addProviderTiles(providers$OpenStreetMap.Mapnik,
group = "OpenStreetMap") %>%
addProviderTiles(providers$Esri.WorldImagery,
group = "Esri World Imagery") %>%
# Chicago boundary
addPolylines(
data = boundary_web,
color = "#1a4f72",
weight = 2,
dashArray = "5,5",
fill = FALSE,
group = "Chicago Boundary",
label = ~NAME
) %>%
# Children Under 5
addPolygons(
data = acs_web,
fillColor = ~pal_children(pct_children_under5),
fillOpacity = 0.8,
color = "grey60",
weight = 0.5,
smoothFactor = 0.2,
popup = popup_children,
highlightOptions = highlightOptions(
weight = 2, color = "#333333", bringToFront = TRUE),
group = "Children Under 5"
) %>%
# Poverty Rate
addPolygons(
data = acs_web,
fillColor = ~pal_poverty(pct_below_poverty),
fillOpacity = 0.8,
color = "grey60",
weight = 0.5,
smoothFactor = 0.2,
popup = popup_poverty,
highlightOptions = highlightOptions(
weight = 2, color = "#333333", bringToFront = TRUE),
group = "Poverty Rate"
) %>%
# PM2.5 Source Coverage
addPolygons(
data = acs_web,
fillColor = "#f4a582",
fillOpacity = 0.65,
color = "grey60",
weight = 0.5,
smoothFactor = 0.2,
popup = popup_pm25_source,
highlightOptions = highlightOptions(
weight = 2, color = "#333333", bringToFront = TRUE),
group = "PM2.5 Source Coverage"
) %>%
# PM2.5 Surface (IDW)
addPolygons(
data = acs_web,
fillColor = ~pal_pm25_surface(pm25_final),
fillOpacity = 0.85,
color = "grey60",
weight = 0.5,
smoothFactor = 0.2,
popup = popup_pm25_surface,
highlightOptions = highlightOptions(
weight = 2, color = "#333333", bringToFront = TRUE),
group = "PM2.5 Surface (IDW)"
) %>%
# Environmental Vulnerability
addPolygons(
data = acs_web,
fillColor = ~pal_vulnerability(env_vulnerability),
fillOpacity = 0.85,
color = "grey60",
weight = 0.5,
smoothFactor = 0.2,
popup = popup_vulnerability,
highlightOptions = highlightOptions(
weight = 2, color = "#333333", bringToFront = TRUE),
group = "Environmental Vulnerability"
) %>%
# Monitor Stations LAST = always on top
addCircleMarkers(
data = aq_points_web,
radius = ~scales::rescale(pm25_mean, to = c(6, 18)),
fillColor = ~pal_monitor(pm25_mean),
fillOpacity = 0.9,
color = "white",
weight = 1.5,
popup = popup_monitors,
label = ~paste0(label_name, ": ", round(pm25_mean, 1), " µg/m³"),
group = "Monitor Stations"
) %>%
# PM2.5 legend — bottomright
addLegend(
position = "bottomright",
pal = pal_monitor,
values = aq_points_web$pm25_mean,
title = "PM2.5 (µg/m³)",
opacity = 0.9
) %>%
# EPA AQI box — bottomleft
addControl(
html = "
<div style='background:white; padding:10px; border-radius:6px;
box-shadow:0 1px 4px rgba(0,0,0,0.2);
font-family:sans-serif; font-size:12px; line-height:1.8'>
<b>EPA Air Quality Index</b><br>
🟢 Good: 0–12 µg/m³<br>
🟡 Moderate: 12–35 µg/m³<br>
🟠 Sensitive Groups: 35–55 µg/m³<br>
🔴 Unhealthy: 55+ µg/m³
</div>",
position = "bottomleft"
) %>%
# Layer control
addLayersControl(
baseGroups = c(
"CartoDB Positron",
"OpenStreetMap",
"Esri World Imagery"
),
overlayGroups = c(
"Chicago Boundary",
"Monitor Stations",
"Children Under 5",
"Poverty Rate",
"PM2.5 Source Coverage",
"PM2.5 Surface (IDW)",
"Environmental Vulnerability"
),
options = layersControlOptions(collapsed = FALSE)
) %>%
# Show PM2.5 Surface and Monitors by default
# Hide everything else
hideGroup(c(
"Children Under 5",
"Poverty Rate",
"PM2.5 Source Coverage",
"Environmental Vulnerability"
)) %>%
# Scale bar topleft
addScaleBar(position = "topleft") %>%
addFullscreenControl() %>%
addResetMapButton() %>%
setView(lng = -87.70, lat = 41.85, zoom = 10.8)This study presents a reproducible, spatially integrated framework for identifying child environmental health vulnerability at the census tract level in Cook County, Chicago. It supports CEHI’s mission to foster environments where all people can prosper, with a focus on addressing children’s unique vulnerabilities through spatially informed research and social justice–driven analysis.
The results show a clear spatial convergence of risk in Chicago’s South and West Sides, where higher poverty rates and greater concentrations of children under five coincide with elevated PM₂.₅ exposure. This pattern underscores how environmental and social stressors compound, demonstrating the interaction of child vulnerability, air pollution, and poverty within specific geographic areas.
Integrating ACS demographic indicators with OpenAQ monitoring data creates a fully spatially referenced data architecture, enabling joint evaluation of multiple determinants of child health. Using IDW interpolation extends exposure estimates beyond monitored locations, supporting continuous and equitable spatial assessment.
CEHI's Research and PracticeThis framework contributes to CEHI’s interdisciplinary approach by:
Identifying priority zones for targeted intervention and field-based sampling
Quantifying inequities through a tract-level vulnerability index grounded in multiple risk dimensions
Enhancing accessibility via an interactive web map that translates complex spatial relationships into actionable insights for stakeholders
Integrating environmental exposure, demographic sensitivity, and spatial inequality within a unified GIS framework, this work advances a data-driven approach to pediatric environmental health. It offers a scalable, transparent foundation for equity-focused decision-making, ensuring interventions target communities where children face the highest combined risk.