This report analyzes traffic accident data in Arizona to explore patterns in fatality rates by county and the relationship between accident severity and weather conditions.
Graph 1 (Fatalities by County): The original graph used a gradient color scheme, which made it hard to differentiate between counties with similar fatality numbers. It was visually confusing and hard to interpret. To fix this, we switched to a categorical color palette that more clearly separates the counties, making it easier to identify those with the highest fatalities.
Graph 2 (Accident Severity by Weather): The original donut chart used an outer ring that was fully connected all the way around, which gave a misleading sense of completeness and made the distribution of accident severity under different weather conditions unclear. We redesigned the chart to show exact proportions of each category, giving a clearer picture of how severity is distributed across weather types.
Question 1: We created a bar graph to show which counties in Arizona have the highest number of traffic fatalities. Maricopa County stood out with the highest number, followed by Pima, Pinal, Mohave, and Yavapai.
Question 2: We visualized accident severity across different weather conditions and found that most accidents occur during clear weather. This suggests that weather is not a strong contributing factor to accident severity. A possible explanation is that drivers may be less cautious when conditions are clear, which ironically leads to more accidents.
Connection: To connect the two questions, we first looked at which counties in Arizona have the highest traffic fatalities (Question 1) to identify where accidents are most deadly. Then, we explored how weather affects accident severity (Question 2) to understand whether weather conditions play a role in making certain areas more dangerous.
By linking these two, we aimed to see if regions with high fatality rates also experience more severe accidents in poor weather conditions, which could suggest weather as a contributing factor to higher fatalities.
To investigate, we created a stacked grouped bar chart that examined whether specific weather conditions affect accident severity in Arizona counties. Our findings showed that weather does not significantly impact accident severity, as most accidents in Maricopa County, the county with the highest fatalities, occurred on clear days. This pattern held true in other counties with high fatality numbers as well like Pima, Pinal, Mohave, and Yavapai, where most accidents also occurred under clear weather.
This suggests that drivers may be less cautious during clear weather, potentially due to a false sense of safety. Rather than weather being a direct cause of higher accident severity, driver behavior in good conditions may be a more influential factor.
Contrast: We used a clearer categorical color palette in place of gradients to help viewers distinguish data more easily. The redesigned donut chart also used better contrast between segments to show proportions clearly.
Repetition: We maintained a consistent color scheme, font style, and layout across all graphics to ensure a unified and professional appearance.
Alignment: Titles, labels, and axes were consistently aligned and formatted to make the graphs visually structured and easy to read.
Proximity: Related elements like labels, legends, and data points were grouped closely together to help viewers quickly interpret the relationships between variables.
We followed Kieran Healy’s principles by focusing on clarity and simplicity. All visualizations were stripped of unnecessary clutter and designed to highlight the core message of the data. We also applied Alberto Cairo’s five qualities of great visualizations:
Beautiful: The charts were designed with clean lines and appropriate color use for a visually pleasing result.
Insightful: The visuals reveal patterns in accident frequency and severity that help us better understand road safety in Arizona.
Enlightening: These insights can help traffic safety officials and policymakers focus efforts where they are most needed.
Truthful: We presented accurate data with proper scaling and no distortion. Misleading color gradients or proportion distortions were corrected.
Functional: The graphs are easy to interpret, with clear legends and meaningful groupings to enhance understanding.
This analysis reveals that Maricopa County has the highest number of traffic fatalities, and clear weather is the most common condition during accidents across all counties. Despite common assumptions, poor weather does not appear to be a major factor in accident severity.
These findings emphasize the importance of driver behavior in good weather conditions, where overconfidence or distraction may lead to more severe outcomes.
Lecture4 (class slides) Lecture5 Lecture 8. Amounts and proportions (D2L class slides) Interactivity 1 - Making Interactive graphs (D2L class slides) Interactivity 2 - Dashboards (D2L class slides) Space (D2L class slides) Geospatial (D2L zip file)
# Load required packages
library(tidyverse)## ── Attaching core tidyverse packages ──────────────────────── tidyverse 2.0.0 ──
## ✔ dplyr 1.1.4 ✔ readr 2.1.5
## ✔ forcats 1.0.0 ✔ stringr 1.5.1
## ✔ ggplot2 3.5.1 ✔ tibble 3.2.1
## ✔ lubridate 1.9.3 ✔ tidyr 1.3.1
## ✔ purrr 1.0.2
## ── Conflicts ────────────────────────────────────────── tidyverse_conflicts() ──
## ✖ dplyr::filter() masks stats::filter()
## ✖ dplyr::lag() masks stats::lag()
## ℹ Use the conflicted package (<http://conflicted.r-lib.org/>) to force all conflicts to become errorslibrary(sunburstR)
library(plotly) ##
## Attaching package: 'plotly'
##
## The following object is masked from 'package:ggplot2':
##
## last_plot
##
## The following object is masked from 'package:stats':
##
## filter
##
## The following object is masked from 'package:graphics':
##
## layoutlibrary(sf)## Linking to GEOS 3.13.0, GDAL 3.8.5, PROJ 9.5.1; sf_use_s2() is TRUElibrary(htmlwidgets)
library(htmltools)
# Load dataset
accident <- read_csv("accident.csv", show_col_types = FALSE)
colnames(accident)## [1] "index" "accident_id" "ST_CASE" "VE_TOTAL"
## [5] "VE_FORMS" "PEDS" "PERSONS" "COUNTY"
## [9] "county_name" "CITY" "city_name" "DAY"
## [13] "MONTH" "YEAR" "HOUR" "MINUTE"
## [17] "NHS" "FUNC_SYS" "func_sys_lit" "ROAD_FNC"
## [21] "road_fnc_lit" "RD_OWNER" "rd_owner_lit" "TWAY_ID"
## [25] "TWAY_ID2" "LATITUDE" "LONGITUD" "SP_JUR"
## [29] "sp_jur_lit" "HARM_EV" "harm_ev_lit" "MAN_COLL"
## [33] "man_coll_lit" "RELJCT1" "RELJCT2" "TYP_INT"
## [37] "WRK_ZONE" "REL_ROAD" "LGT_COND" "lgt_cond_lit"
## [41] "WEATHER" "weather_lit" "SCH_BUS" "CF1"
## [45] "CF2" "CF3" "cf1_lit" "cf2_lit"
## [49] "cf3_lit" "FATALS" "A_INTER" "a_inter_lit"
## [53] "A_ROADFC" "a_road_fc_lit" "A_TOD" "a_tod_lit"
## [57] "A_DOW" "a_dow_lit" "A_LT" "a_lt_lit"
## [61] "A_SPCRA" "a_spcra_lit" "A_PED" "a_ped_lit"
## [65] "A_PED_F" "a_ped_f_lit" "A_PEDAL" "a_pedal_lit"
## [69] "A_PEDAL_F" "a_pedal_f_lit" "A_POLPUR" "a_polour_lit"
## [73] "A_POSBAC" "a_posbac_lit" "A_DIST" "a_dist_lit"
## [77] "A_DROWSY" "a_drowsy_lit" "INDIAN_RES" "indian_res_lit"# Summarize accident data
county_fatal <- accident %>%
filter(!is.na(county_name), !is.na(FATALS)) %>%
group_by(county_name = tolower(county_name)) %>%
summarise(total_fatalities = sum(FATALS), .groups = "drop")
# Load shapefile
us_counties <- read_sf("cb_2022_us_county_5m/cb_2022_us_county_5m.shp")
colnames(us_counties)## [1] "STATEFP" "COUNTYFP" "COUNTYNS" "AFFGEOID" "GEOID"
## [6] "NAME" "NAMELSAD" "STUSPS" "STATE_NAME" "LSAD"
## [11] "ALAND" "AWATER" "geometry"# filter for Arizona
az_counties <- us_counties %>%
filter(STUSPS == "AZ") %>%
mutate(NAME = tolower(NAME))
# Join and handle missing data
az_map_data <- az_counties %>%
left_join(county_fatal, by = c("NAME" = "county_name")) %>%
mutate(total_fatalities = replace_na(total_fatalities, 0))# Categorize fatalities
az_map_data <- az_map_data %>%
mutate(
fatality_category = cut(
total_fatalities,
breaks = c(-1, 20, 40, 60, 80, 100, 120, 140, 160, 180, 241, Inf),
labels = c("0-20", "21-40", "41-60", "61-80", "81-100", "101-120", "121-140", "141-160", "161-180", "181-240", "241+"),
include.lowest = TRUE,
right = TRUE
),
fatality_category = factor(
fatality_category,
levels = c("0-20", "21-40", "41-60", "61-80", "81-100", "101-120", "121-140", "141-160", "161-180", "181-240", "241+")
),
tooltip = paste0("County: ", NAME, "\nFatalities: ", total_fatalities)
)
# Plot
map <- ggplot(az_map_data) +
geom_sf(aes(fill = fatality_category, text = tooltip), color = "white", linewidth = 0.2) +
scale_fill_viridis_d(option = "plasma", na.value = "grey90") +
labs(
title = "Total Accident Fatalities by County in Arizona",
fill = "Fatality Range",
x = "Longitude",
y = "Latitude"
) +
coord_sf(crs = st_crs("EPSG:26948")) +
theme_minimal() +
theme(
plot.title = element_text(hjust = 0.5),
axis.title.x = element_text(size = 12),
axis.title.y = element_text(size = 12),
axis.text = element_text(size = 10),
axis.text.x = element_text(angle = 40, hjust = 1)
)
ggplotly(map, tooltip = "text")# Define a colorblind-friendly palette
color_palette <- c(
"#7AD151", # 0–20
"#F0F921", # 41–60
"#FDC326", # 61–80
"#F89D45", # 81–100
"#E76F5B", # 101–120
"#D1497F", # 121–140
"#AA2395", # 141–160
"#7E03A8", # 161–180
"#0D0887" # 241+
)
# Create severity group based on FATALS
accident_clean <- accident %>%
filter(!is.na(weather_lit), !is.na(FATALS)) %>%
mutate(SeverityGroup = case_when(
FATALS == 0 ~ "No Fatalities",
FATALS == 1 ~ "1 Fatality",
FATALS >= 2 ~ "2+ Fatalities"
))
# Create sunburst path with two layers
weather_fatal <- accident_clean %>%
count(weather_lit, SeverityGroup) %>%
mutate(
path = paste(weather_lit, SeverityGroup, sep = "-") # Create hierarchical path
) %>%
select(path, n)
# Prepare data for sunburst
weather_fatal_hierarchy <- accident_clean %>%
count(weather_lit) %>%
mutate(path = weather_lit) %>%
select(path, n) %>%
bind_rows(weather_fatal)
# Plot
sunburst_plot <- sunburst(weather_fatal_hierarchy, colors = color_palette)
prependContent(
sunburst_plot,
tags$div(
style = "text-align:center; font-weight:normal; font-size:22px;
font-family: 'Helvetica Neue', Helvetica, Arial, sans-serif;
margin-bottom:10px;",
"Accident Severity by Weather Condition"
)
)# Group and summarize data
county_weather_severity <- accident %>%
filter(!is.na(county_name), !is.na(weather_lit), !is.na(FATALS)) %>%
mutate(
county_name = tolower(county_name),
severity = case_when(
FATALS == 0 ~ "No Fatality",
FATALS == 1 ~ "1 Fatality",
FATALS >= 2 ~ "2+ Fatalities"
)
) %>%
group_by(county_name, weather_lit, severity) %>%
summarise(total_fatalities = sum(FATALS), .groups = "drop")
# Plot
barchart <- ggplot(county_weather_severity, aes(
x = reorder(county_name, -total_fatalities),
y = total_fatalities,
fill = severity,
text = paste0("County: ", county_name, "\nFatalities: ", total_fatalities, "\nSeverity: ", severity)
)) +
geom_bar(stat = "identity") +
facet_wrap(~ weather_lit, scales = "free_x") +
scale_fill_viridis_d(option = "plasma") +
labs(
title = "Accident Fatalities by County and Weather Condition",
x = "County",
y = "Number of Fatalities",
fill = "Severity Level"
) +
theme_minimal() +
theme(
plot.title = element_text(hjust = 0.5),
axis.text.x = element_text(angle = 45, hjust = 1),
legend.position = "bottom"
)
ggplotly(barchart, tooltip = "text")