Code
/* Properly place and align figure captions */
div.leaflet-control {
text-align: left;
}Food Service Inspection Analysis
Montgomery County, Maryland
Introduction
Code
# Configure working directory
setwd('~/Library/CloudStorage/OneDrive-Personal/Loyola/DS_736/GIT_Repository/DS736')
# Load libraries
library(plyr)
library(httr)
library(jsonlite)
library(lubridate)
library(tidyverse)
library(scales)
library(ggrepel)
library(leaflet)
library(sf)
library(plotly)
library(patchwork)
library(maps)
# Use parameter from header to determine data source
# If data refresh is indicated, pull data from dataMontgomery and store in data table
if (params$refresh_data) {
data = GET('https://data.montgomerycountymd.gov/resource/46p5-g5na.json?$limit=15000')
data = as_tibble(fromJSON(rawToChar(data$content)))
} else {
# Otherwise use data already downloaded
data <- as_tibble(read_csv('MC_FoodInspection.csv'))
}
# Clean up data
data <- data |>
# Use descriptive titles for violations
rename(
v_Approved_Source = violation1,
v_Contamination_Protection = violation2,
v_Workers_Restricted = violation3,
v_Hand_Washing = violation4,
v_Cooling_Time_Temp = violation5,
v_Cold_Holding_Temp = violation6a,
v_Hot_Holding_Temp = violation6b,
v_Cooking_Time_Temp = violation7a,
v_Reheat_Time_Temp = violation7b,
v_Hot_Cold_Water = violation8,
v_Sewage_Disposal = violation9,
v_Toxic_Substances = violation20,
v_Pest_Control = violation22,
v_Nutrition_Labeling = nutritional_labeling,
v_Trans_Fat = violationtransfat,
v_No_Smoking = violationsmoking
) |>
# Remove unused columns
select(-c(address2, type)) |>
# Remove observations with NA values where important
filter(!is.na(inspectiontype)) |>
# Convert NA values in category to 'Other'
mutate(category = replace_na(category, 'Other')) |>
# Convert inspection date to date format
mutate(inspectiondate = as_date(inspectiondate)) |>
# Convert other columns to factors
mutate_at(.vars = c(3:4, 6:23, 24:25), as_factor) |>
# Convert Lat / Long to numeric
mutate(
latitude = as.numeric(latitude),
longitude = as.numeric(longitude)
)Montgomery County is located to the north and west of Washington, DC in the state of Maryland (see Figure 1). Montgomery County occupies roughly 493 square miles, and, as of the 2020 census, had a population 1,062,061 individuals (U.S. Census Bureau QuickFacts, n.d.). The Department of Health and Human Services, the largest department within the Montgomery County Government, is responsible for protecting community health and addressing basic human needs across the county (Montgomery County Department of Health and Human Services – Home Page – About Us, n.d.). Services provided by the department include the licensing and inspection of food service facilities such as grocery stores, restaurants, cafeterias, and food trucks (Licensure and Regulatory Services - Montgomery County, Maryland, n.d.).
This report presents the analysis of data provided by the Department of Health and Human Services regarding the routine inspection of food service facilities across Montgomery County for the most recent 24 months.
Code
# Create map to show where Montgomery County is located within the state of Maryland
maryland <- map_data('county') |>
filter(region == 'maryland') |>
mutate(highlight = (subregion == 'montgomery'))
# Find the center of Montgomery County to guide label positioning
mc_center <- maryland |>
filter(subregion == 'montgomery') |>
summarize(lat = mean(lat), long = mean(long))
# Create the map with a label for Montgomery County
maryland |>
ggplot(
aes(
x = long,
y = lat
)
) +
geom_polygon(
aes(
group = group,
fill = highlight
),
color = 'black'
) +
theme_void() +
theme(legend.position = 'none') +
geom_label(
aes(
x = mc_center$long,
y = mc_center$lat,
label = 'Montgomery County'
),
nudge_x = -0.5,
nudge_y = -0.05
)
Data Set
The Food Inspection data set is provided by the Open Data Operations of the Montgomery County Government (“Montgomery County Data Open Data Portal,” n.d.). Contained within the data set are the results of 12,985 food service facility inspections conducted during the most recent 24 months. For each inspection, the following information is provided:
| Variable | Description |
|---|---|
| Name | Facility name |
| Address | Facility street address |
| City | Facility city |
| Zip | Facility zip code |
| Date | Inspection date |
| Result | Overall inspection result |
| Approved Source | All food from a regulated source |
| Contamination Protection | All food protected from contamination |
| Ill Worker Restriction | Food not handled by worker with transmissible disease |
| Hand Washing | Workers follow proper hand washing regulations |
| Hot Holding Temperature | Hot foods held at or above proper temperature |
| Cooking Time and Temperature | Foods cooked to an appropriate internal temperature |
| Reheat Time and Temperature | Foods are reheated to an appropraite internal temperature |
| Hot and Cold Water | Both hot and cold running water are available |
| Proper Sewage Disposal | All sewage is disposed properly |
| Toxic Substances | Toxic substances are properly marked and stored |
| Rodents and Insects | Proper measures are taken to prevent and eliminate rodents and insects |
| Nutrition Labeling | Facility follows nutritional labeling regulations |
| Trans Fat | Facility adheres to restrictions regarding trans fat |
| No Smoking Sign | No smoking signs posted per regulations |
| Inspection Type | Comprehensive or monitoring inspection |
| Owner | Facility owner |
| Category | Facility cataegory |
| Type | Facility type |
| Lat | Latitude of facility |
| Long | Longitude of facility |
Findings
Code
# Create a list of likely chains in data set for further cleaning
chains <- c(
'STARBUCKS',
'7-ELEVEN',
'ADVANCE AUTO PARTS',
'ALDI',
'AMAZON FRESH',
'AMERICAN LEGION',
'AUTUMN LAKE HEALTHCARE',
'BURGER KING',
'CALIFORNIA TORTILLA',
'CAVA MEZZE GRILL',
'CHICK-FIL-A',
'CHIPOTLE',
'CHOPT CREATIVE SALAD',
'CVS',
'DOLLAR TREE',
'DOMINO\\\'S PIZZA',
'DON POLLO',
'DUNKIN DONUTS',
'EINSTEIN BROS. BAGELS',
'FIVE BELOW',
'FIVE GUYS',
'GENERAL NUTRITION CENTER',
'GENJI SUSHI',
'GIANT FOOD',
'GOLDS GYM',
'HARRIS TEETER',
'HISSHO SUSHI',
'HOME DEPOT',
'HOME GOODS',
'IHOP',
'LEDO PIZZA',
'LITTLE CAESARS',
'MAMMA LUCIA',
'MCDONALD',
'MEGAMART',
'MICHAEL\\\'S STORE',
'MISSION BBQ',
'OLD NAVY',
'PANDA EXPRESS',
'PANERA BREAD',
'PAPA JOHN\\\'S PIZZA',
'PARTY CITY',
'PIZZA HUT',
'POTBELLY',
'PROMEDICA',
'RED LOBSTER',
'RITA\\\'S',
'ROSS DRESS FOR LESS',
'SAFEWAY',
'STAPLES',
'SUBWAY',
'TARGET',
'TGI FRIDAY\\\'S',
'T.J. MAXX',
'TACO BELL',
'TRADER JOE\\\'S',
'VITAMIN SHOPPE',
'WALGREEN',
'WENDY\\\'S',
'WHOLE FOODS MARKET',
'WINGSTOP'
)
# Get count of locations by name (i.e. if > 1, likely a chain) then select top 10
data <- data |>
# Clean up names of chain establishments based on list above
# e.g. STARBUCKS 123 and STARBUCKS KIOSK both become STARBUCKS
mutate(
name = str_replace(
name,
paste(
'[[:graph:]|[:space:]]*(',
paste(chains, collapse = '|'),
')[[:graph:]|[:space:]]*',
sep = ''
),
'\\1'
)
) |>
mutate(
name = as_factor(
ifelse(
name =='CVS',
name,
ifelse(
name == 'MCDONALD',
'McDonald\'s',
str_to_title(name)
)
)
)
)
# Determine how many unique food service facilities are in the database
distinct_locations <- data |>
distinct(name, latitude, longitude, category, .keep_all = TRUE)Across the 4,103 unique food service facilities identified in the data set, the Montgomery County Department of Health conducted conducted 12,985 inspections during the most recent 24 months. This analysis considers the types of facilities, the quantity of inspections conducted, types of violations recorded, and facility closures resulting from a failed inspection. Each of these considerations is presented in a separate tab below.
The 4,103 unique food service facilities in the data set are broken down by category. The top categories based on number of facilities are presented in Figure 2, with all remaining categories lumped into the designation Other.
Code
# Group public schools together
distinct_locations <- distinct_locations |>
dplyr::mutate(
category = str_replace(
category,
'^(Public School).*',
'\\1'
)
)
# Find the top categories by count, excluding 'Other'
top_categories <- distinct_locations |>
filter(category != 'Other') |>
group_by(category) |>
summarize(n = length(category)) |>
arrange(desc(n)) |>
slice(1:7)
# Group all remaining categories into 'Other'
topN_categories <- distinct_locations |>
mutate(
category = ifelse(category %in% top_categories$category, category, 'Other')
) |>
group_by(category) |>
summarize(n = length(category)) |>
mutate(
category = factor(
category,
levels = c(top_categories$category, 'Other')
)
)
# Plot the top categories of food service facilities
topN_categories |>
arrange(fct_relevel(category, 'Other', after = Inf)) |>
plot_ly(
labels = ~category,
values = ~n,
textposition = 'inside'
) |>
add_pie(
title = paste0('Total Facilities\n', comma(sum(topN_categories$n))),
hole = 0.7,
direction = 'clockwise',
sort = FALSE,
hovertemplate = '%{label}<br>Percent: %{percent}<br>Count: %{value:,}<extra></extra>',
)Code
# Identify likely chains
locations_count <- distinct_locations |>
group_by(name) |>
# Count the number of unique locations for each establishment
summarize(
locations = length(name),
) |>
# Order by count from largest to smallest and select the top 10
arrange(desc(locations))
# Count the number of locations likely part of a chain versus
# those not part of a chain (independents)
independents <- sum(locations_count$locations == 1)
chain_count = nrow(distinct_locations) - independentsAs indicated in Figure 2, the majority of food service facilities in the data set, are categorized as Restaurants and Markets. Many of those restaurants and markets are part of a chain, with approximately 1,064 facilities in the data set belonging to a food service chain. In total, there appear to be approximately 211 different chains represented in the data set. The 10 largest chains, based on number of locations, are identified in Figure 3.
Code
# Plot the top 10 chains in Montgomery County
locations_count |>
slice(1:10) |>
ggplot(aes(x = locations, y = fct_reorder(name, locations))) +
# geom_col(aes(fill = fct_reorder(name, locations))) +
geom_col(aes(fill = locations)) +
geom_text(aes(label = locations), hjust = -0.25) +
labs(
# title = 'Figure 1: 10 Largest Food Service Chains in Montgomery County',
x = 'Locations',
y = NULL,
caption = 'Data from data.montgomerycountymd.gov'
) +
theme_light() +
theme(
legend.position = 'none',
plot.title = element_text(hjust = 0.5)
) +
scale_fill_continuous(type = 'viridis', trans = 'reverse') +
scale_x_continuous(
labels = seq(0, round_any(max(locations_count$locations), 10, ceiling), 10),
breaks = seq(0, round_any(max(locations_count$locations), 10, ceiling), 10)
)
Code
# Group public schools together into one category
data <- data |>
mutate(
category = str_replace(
category,
'^(Public School).*',
'\\1'
)
)
# Determine the number of inspections per week in the data set
inspections_per_week <- data |>
mutate(
Year = as_factor(year(inspectiondate)),
Week = strftime(inspectiondate, format = '%V')
) |>
group_by(Year, Week) |>
summarize(n = length(inspectiondate), .groups = 'drop')The Montgomery County Department of Health recorded 12,985 inspections in the provided data set. A histogram of the number of inspections conducted per week is presented in Figure 4. During the past 24 months, the Department of Health conducted an average of 123.7 inspections per week, with a standard deviation of 44.5. The coefficient of variation of 36% indicates moderate variation in the number of inspections conducted per week.
Code
# Get mean and standard deviation to further augment the graph
stat_marker <- tribble(
~stat, ~value,
'Mean', mean(inspections_per_week$n),
'+1 SD', mean(inspections_per_week$n) + sd(inspections_per_week$n),
'-1 SD', mean(inspections_per_week$n) - sd(inspections_per_week$n)
) |>
mutate(stat = factor(stat))
# Create a histogram of the number of inspections per week.
# Highlight mean and ±1 standard deviations
inspections_per_week |>
ggplot(aes(x = n)) +
geom_histogram(
# aes(y = after_stat(density)),
binwidth = 10,
fill = 'deepskyblue4'
) +
labs(
x = 'Inspections per Week',
y = 'Count',
caption = 'Data from data.montgomerycountymd.gov'
) +
geom_vline(
data = stat_marker,
aes(xintercept = value, color = stat),
linetype = 'dashed'
) +
scale_color_manual(
guide = 'none',
# '',
values = c('darkorange1', 'darkorange1', 'darkolivegreen3'),
# labels = c('+1 SD', 'Mean', '-1 SD'),
# breaks = c('+1 SD', 'Mean', '-1 SD')
) +
theme_minimal() +
geom_point(
data = stat_marker,
aes(x = value, y = 0, color = stat),
size = 3
) +
geom_label_repel(
data = stat_marker,
aes(
x = value,
y = 0,
label = paste0(stat, ': ', round(value, 1))
),
box.padding = 1,
point.padding = 1,
size = 4,
color = 'Grey50',
segment.color = 'white'
)
Figure 5 breaks down the number of inspections by month for the past 24 months. From the results, we note that the number of inspections seem to peak in March and tend to be lowest in the summer months. Note that partial months are not included in the graph.
Code
# Get number of inspections by month
# Ignore partial months
today <- Sys.Date()
end_date <- floor_date(today, 'month') - days(1)
start_date <- end_date - years(2) + days(1)
start_date <- start_date + months(1)
# Get count of inspections for past 24 months (ignoring partial months)
inspection_count <- data |>
select(inspectiondate) |>
filter(inspectiondate >= start_date & inspectiondate <= end_date) |>
mutate(
Month = as_factor(month(inspectiondate, label = TRUE, abbr = TRUE)),
Year = as_factor(year(inspectiondate))
) |>
group_by(Year, Month) |>
summarize(Inspections = length(Month), .groups = 'keep')
# Plot number of inspections by month
inspection_count |>
ggplot(aes(x = Month, y = Inspections)) +
geom_line(
aes(group = Year, color = Year),
linewidth = 2
) +
geom_point(
aes(fill = Year),
size = 5,
shape = 21,
color = 'white'
) +
labs(
x = 'Month',
y = 'Inspections Conducted',
caption = paste0('Data as of ', today ,'\nPartial months not displayed')
) +
scale_color_brewer(palette = 'Dark2') +
scale_fill_brewer(palette = 'Dark2') +
scale_y_continuous(
expand = c(0, 0),
limits = c(0, round_any(max(inspection_count$Inspections), 200, ceiling)),
labels = seq(0, round_any(max(inspection_count$Inspections), 200, ceiling), 200),
breaks = seq(0, round_any(max(inspection_count$Inspections), 200, ceiling), 200)
) +
theme_light() +
theme(plot.title = element_text(hjust = 0.5)) +
geom_label_repel(
aes(label = ifelse(
Inspections == min(Inspections) | Inspections == max(Inspections),
comma(Inspections), '')),
box.padding = 1,
point.padding = 1,
size = 4,
color = 'Grey50',
segment.color = 'darkblue'
)
There are 16 catgories of violations that may be noted during a food service facility inspection. The four most common violations, as a percentage of inspections, are identifed in Figure 6.
Code
# Create descriptive labels for violations for graphing
violations = c(
'v_Approved_Source' = 'Food From\nApproved Sources',
'v_Contamination_Protection' = 'No Contaminated\nFoods',
'v_Workers_Restricted' = 'No Workers with\nTransmissible Disease',
'v_Hand_Washing' = 'Proper\nHand Washing',
'v_Cooling_Time_Temp' = 'Cooling Time\nand Temperature',
'v_Cold_Holding_Temp' = 'Cold Holding\nTemperature',
'v_Hot_Holding_Temp' = 'Hot Holding\nTemperature',
'v_Cooking_Time_Temp' = 'Cooking Time\nand Temperature',
'v_Reheat_Time_Temp' = 'Reheat Time\nand Temperature',
'v_Hot_Cold_Water' = 'Hot and Cold\nWater Available',
'v_Sewage_Disposal' = 'Proper Sewage\nDisposal',
'v_Toxic_Substances' = 'Toxic Substances\nControl',
'v_Pest_Control' = 'Pest Control\nMeasures',
'v_Nutrition_Labeling' = 'Nutritional\nLabeling',
'v_Trans_Fat' = 'Trans Fats',
'v_No_Smoking' = 'No-Smoking\nSigns'
)
# Calculate percent of violations noted based on number of inspections
violations_pct <- data |>
pivot_longer(cols = 7:22, names_to = 'Violation', values_to = 'Result') |>
count(Violation, Result) |>
group_by(Violation) |>
mutate(pct = n / sum(n) * 100)
# Record the four most common violations by percentage
top_violations <- violations_pct |>
arrange(desc(Result), desc(pct)) %>%
filter(Result == 'Out of Compliance') |>
ungroup() |>
slice(1:4)
# Plot donut charts of the four most common violations
plot_ly(
labels = ~Result,
values = ~n,
textposition = 'inside'
) |>
add_pie(
data = violations_pct[violations_pct$Violation == top_violations$Violation[[1]], ],
title = violations[top_violations$Violation[[1]]],
hole = 0.7,
hovertemplate = '%{label}<br>Percent: %{percent}<br>Count: %{value:,}<extra></extra>',
domain = list(
row = 0,
column = 0
)
) |>
add_pie(
data = violations_pct[violations_pct$Violation == top_violations$Violation[[2]], ],
title = violations[top_violations$Violation[[2]]],
hole = 0.7,
hovertemplate = '%{label}<br>Percent: %{percent}<br>Count: %{value:,}<extra></extra>',
domain = list(
row = 0,
column = 1
)
) |>
add_pie(
data = violations_pct[violations_pct$Violation == top_violations$Violation[[3]], ],
title = violations[top_violations$Violation[[3]]],
hole = 0.7,
hovertemplate = '%{label}<br>Percent: %{percent}<br>Count: %{value:,}<extra></extra>',
domain = list(
row = 1,
column = 0
)
) |>
add_pie(
data = violations_pct[violations_pct$Violation == top_violations$Violation[[4]], ],
title = violations[top_violations$Violation[[4]]],
hole = 0.7,
hovertemplate = '%{label}<br>Percent: %{percent}<br>Count: %{value:,}<extra></extra>',
domain = list(
row = 1,
column = 1
)
) |>
layout(
showlegend = TRUE,
grid = list(rows = 2, columns = 2)
)The diagram in Figure 7 highlights the most common violations, as a percentage of inspections, by food service facility categories. From this diagram, it appears that the most common violations, especially those identified in Figure 6, tend to occur predominantly in hospitals, restaurants, and nursing homes. (Note: 0.0% indicates no occurrence of the specific violation was noted during any of the inspections for the given food service facility category.)
Code
# Find the top categories of food service providers based on percentage of
# violations across all violations recorded, along with the top violations recorded
# First find the percentage of violations by category
violations_by_category <- data |>
dplyr::mutate(
category = str_replace(
category,
'^(Public School).*',
'\\1'
)
) |>
pivot_longer(cols = 7:22, names_to = 'violation', values_to = 'result') |>
select(category, violation, result) |>
count(category, violation, result) |>
group_by(category, violation) |>
mutate(pct = n / sum(n) * 100) |>
ungroup() |>
filter(result == 'Out of Compliance')
# Rank categories based on percentages of noncompliances across violations
top_category <- violations_by_category |>
filter(!category == 'Other') |>
filter(!is.na(category)) |>
group_by(category) |>
summarize(tot_pct = sum(pct)) |>
arrange(desc(tot_pct)) |>
select(category) |>
slice(1:9)
# Rank violations based on percentages of noncompliances across violations
top_violation <- violations_by_category |>
group_by(violation) |>
summarize(tot_pct = sum(pct)) |>
arrange(desc(tot_pct)) |>
select(violation) |>
slice(1:9)
# Create a 10th category called 'Other' to include all other food service
# categories not in the top 9. Also create a 10th category called 'Other'
# for violations not in the top 9.
pct_violation_by_category <- data |>
dplyr::mutate(
category = str_replace(
category,
'^(Public School).*',
'\\1'
)
) |>
pivot_longer(cols = 7:22, names_to = 'violation', values_to = 'result') |>
select(category, violation, result) |>
mutate(
category = ifelse(
category %in% top_category$category,
category,
'Other'
),
category = factor(
category,
levels = c(top_category$category, 'Other')
),
violation = ifelse(
violation %in% top_violation$violation,
violation,
'Other'
),
violation = factor(
violation,
levels = c(top_violation$violation, 'Other')
)
) |>
# Find percentage of violations by inspection for each combination of
# category and violation
count(category, violation, result) |>
group_by(category, violation) |>
mutate(pct = n / sum(n) * 100) |>
ungroup() |>
filter(result == 'Out of Compliance') |>
select(category, violation, pct) |>
pivot_wider(names_from = violation, values_from = pct) |>
mutate_at(.vars = 2:11, ~replace_na(., 0)) |>
mutate_at(.vars = 'category', ~replace_na(., 'Other')) |>
pivot_longer(cols = 2:11, names_to = 'violation', values_to = 'pct') |>
arrange(desc(pct))
# Find the max percentage to set up legend for graph
max_pct <- round_any(max(pct_violation_by_category$pct), 10, ceiling)
# Create a heat map to display percent occurrence of violations by
# category and violation
pct_violation_by_category |>
ggplot(aes(x = category, y = violation, fill = sqrt(pct))) +
geom_tile(color = 'black') +
geom_text(
aes(label = sprintf('%.1f%%', pct)),
# aes(label = paste0(round(pct, 1), '%')),
size = 3,
) +
scale_y_discrete(
limits = rev(c(top_violation$violation, 'Other')),
labels = c(violations[top_violation$violation], 'Other')
) +
theme_minimal() +
theme(axis.text.x = element_text(angle = 50, vjust = 1, hjust = 1)) +
scale_fill_continuous(
low = 'white',
high = 'red',
labels = paste0(c(seq(0, max_pct, by = 10)), '%'),
breaks = c(sqrt(seq(0, max_pct, by = 10))),
name = '% Non-Compliant'
) +
labs(
x = NULL,
y = NULL,
caption = 'Data from data.montgomerycountymd.gov'
) +
guides(
fill = guide_legend(
reverse = TRUE,
override.aes = list(color = 'black')
)
)
Based on the data set, it does not appear that all noted violations during an inspection result in a closure of the facility. Figure 8 presents the percentage of inspections that resulted in a facility closure by category, along with the total number of closures by category. While restaurants had the highest number of closures, they also represent the largest category. Carry-out facilities represent a smaller portion of the food service facilities in the data set, but had the highest percentage of inspections resulting in a facility closure.
Code
# Find top categories for which an inspection resulted in a closure
failure_by_category <- data |>
group_by(category) |>
summarize(
n = sum(inspectionresults == 'Facility Closed'),
pct = round(n / length(category), 4)
) |>
arrange(desc(pct)) |>
filter(category != 'Other') |>
mutate(category = as.character(category)) |>
filter(n > 5)
# For all other categories, group as 'Other'
top_failure_by_category <- data |>
mutate(category = as.character(category)) |>
mutate(
category = ifelse(
category %in% failure_by_category$category,
category,
'Other'
)
) |>
mutate(
category = factor(
category,
levels = c(failure_by_category$category, 'Other')
)
) |>
group_by(category) |>
summarize(
n = sum(inspectionresults == 'Facility Closed'),
pct = round(n / length(category), 4),
count = length(category)
)
# Set scale for second axis
p_scale <- 1 / 4000
max_pct <- max(top_failure_by_category$pct)
# Plot total closures and percent closures
top_failure_by_category |>
ggplot(aes(x = category, y = pct)) +
# Use bar chart for percent of closure by category
geom_col(aes(fill = sqrt(pct))) +
geom_text(
aes(
label = percent(pct)
),
hjust = -0.25
) +
# Use line chart for total closures by category
geom_line(
aes(
x = category,
y = n * p_scale,
color = 'Closure\nCount',
group = 1
),
linewidth = 1
) +
geom_point(
aes(
x = category,
y = n * p_scale,
group = 1
),
size = 3,
shape = 21,
fill = 'white',
color = 'black'
) +
coord_flip() +
labs(
x = NULL,
y = 'Percent of Inspections Resulting in Closure',
caption = 'Data from data.montgomerycountymd.gov'
) +
theme_light() +
scale_color_manual(
NULL,
values = 'black'
) +
scale_fill_continuous(
low = 'lightpink',
high = 'red',
labels = paste0(c(100 * seq(0, max_pct, by = 0.005)), '%'),
breaks = c(sqrt(seq(0, max_pct, by = 0.005))),
name = '% Closures',
guide = 'none'
) +
scale_y_continuous(
limits = c(0, round_any(max(top_failure_by_category$pct), .01, ceiling)),
sec.axis = sec_axis(
~. / p_scale,
name = 'Count of Inspections Resulting in Closure'
),
labels = percent
) +
scale_x_discrete(
limits = rev(c(failure_by_category$category, 'Other')),
labels = rev(c(failure_by_category$category, 'Other'))
) +
geom_label_repel(
aes(
x = category,
y = n * p_scale,
label = paste0(n, ' of ', comma(count)),
group = 1
),
nudge_x = 0.01,
box.padding = 1,
point.padding = 1,
size = 4,
color = 'Grey50',
segment.color = 'darkblue'
) +
guides(
color = guide_legend(order = 1),
fill = FALSE
)
The map presented in Figure 9 identifies the location of the food service facilities that were closed in 2023 as a result of an inspection by the Department of Health. The color coding identifies if the facility remains closed at present or has since reopened following a favorable re-inspection. (Note: Map controls in the lower-right corner allow for filtering of facilities based on current status.)
Code
# Identify all inpsection failures for 2023
inspection_failure <- data |>
mutate(Year = as_factor(year(inspectiondate))) |>
filter(Year == 2023) |>
filter(inspectionresults == 'Facility Closed') |>
select(name, latitude, longitude) |>
unique()
# Determine which facilities have since reopened
failure_status <- data |>
filter(
name %in% inspection_failure$name &
latitude %in% inspection_failure$latitude &
longitude %in% inspection_failure$longitude
) |>
group_by(name, latitude, longitude) |>
filter(inspectiondate == max(inspectiondate)) |>
ungroup() |>
select(name, inspectiondate, inspectiontype, inspectionresults, latitude, longitude) |>
mutate(
inspectionresults = as.factor(
ifelse(
inspectionresults == 'Facility Closed',
'Closed',
'Reopened'
)
)
)
# Create color coding for facilities that remain closed versus those that have reopened
pal <- colorFactor(c('red', 'darkgreen'), domain = c('Closed', 'Reopened'))
# Add a shape file to highlight Montgomery County on the map
# This shapefile was downloaded from the US Census Bureau
load(file = 'MC_Shapefile.RData')
# Create a Leaflet map of Montgomery County
leaflet(data = failure_status) |>
addProviderTiles(providers$Esri.WorldTopoMap) |>
addPolygons(
data = counties,
weight = 2,
fillColor = 'lightblue',
fillOpacity = 0.3
) |>
# Add markers to highlight facilities that remain closed
addCircleMarkers(
data = failure_status[failure_status$inspectionresults == 'Closed', ],
lng = ~longitude,
lat = ~latitude,
radius = 5,
color = ~pal(inspectionresults),
fillOpacity = 0.7,
popup = ~paste(
name, '<br/>',
'Last Inspection:', inspectiondate, '<br/>',
'Inspection Type:', inspectiontype, '<br/>',
'Status:', inspectionresults
),
label = ~name,
group = 'Closed'
) |>
# Add markers to highlight facilities that have since reopened
addCircleMarkers(
data = failure_status[failure_status$inspectionresults == 'Reopened', ],
lng = ~longitude,
lat = ~latitude,
radius = 5,
color = ~pal(inspectionresults),
fillOpacity = 0.7,
popup = ~paste(
name, '<br/>',
'Last Inspection:', inspectiondate, '<br/>',
'Inspection Type:', inspectiontype, '<br/>',
'Current Status:', inspectionresults
),
label = ~name,
group = 'Reopened'
) |>
addLegend(
position = 'topright',
pal = pal,
values = ~inspectionresults,
labels = ~c('Closed', 'Reopened'),
title = 'Current Status'
) |>
# Add a label to emphasize 'Montgomery County'
addControl(
'Montgomery County, MD',
position = 'topleft'
) |>
# Add layer control to allow user to declutter map by removing markers
addLayersControl(
overlayGroups = c('Closed', 'Reopened'),
options = layersControlOptions(collapsed = FALSE),
position = 'bottomright'
) |>
addControl(
'Show Status:',
position = 'bottomright'
)Conclusions
The Montgomery County Department of Health conducted 12,985 inspections of food service facilities during the past 24 months. This report highlights several details regarding the inspections, including the types of facilities, inspections conducted, the most common violations, and the types and locations of facilities that were closed as a result of an inspection.