1: Load Dataset

# Load my dataset 
car_accidents <- read.csv("US_Accidents_March23.csv")

2: Number of Accidents by State

# Count accidents by state
state_counts <- table(car_accidents$State)

# Sort in decreasing order
state_counts <- sort(state_counts, decreasing = TRUE)
barplot(state_counts,
        main = "Number of Accidents by State",
        xlab = "State",
        ylab = "Number of Accidents",
        las = 3,         # rotate x-axis labels vertically
        cex.names = .5, # reduce font size
        col = "steelblue")

The bar plot shows which states have the most recorded accidents. California, Texas, and Florida have the highest counts, indicating higher traffic density or reporting.

3: Mean Severity of Accidents

# Calculate mean severity
mean_severity <- mean(car_accidents$Severity, na.rm = TRUE)
mean_severity
## [1] 2.212384

The mean severity (~2.21) indicates that most accidents are moderate in impact. The summary shows the distribution: most accidents are level 2, fewer are level 1 or 4.

4: Correlation Between Temperature and Visibility

# Correlation between Temperature(F) and Visibility(mi)
correlation <- cor(car_accidents$Temperature.F., car_accidents$Visibility.mi., use = "complete.obs")
correlation
## [1] 0.2171733

The correlation shows a slight relationship between temperature and visibility. The scatter plot confirms that higher temperatures have a minor effect on visibility during accidents.

5: Histogram of Temperature During Accidents

# Histogram of Temperature
hist(car_accidents$Temperature.F.,
     breaks = 40,
     main = "Distribution of Temperature During Accidents",
     xlab = "Temperature (F)",
     ylab = "Number of Accidents",
     col = "orange",
     border = "black",
     xlim = c(-50, 150))

Most accidents occur between moderate temperatures, with fewer accidents at extreme low or high temperatures. The histogram shows a roughly normal-like distribution with slight skew.

6: Day vs Night Accidents Severity Comparison (T-Test)
 There is a statistically significant difference in accident severity between day and night. On average, night accidents tend to be slightly more severe than day accidents.

# Divide dataset into Day vs Night accidents
day_accidents <- car_accidents[car_accidents$Sunrise_Sunset == "Day", ]
night_accidents <- car_accidents[car_accidents$Sunrise_Sunset == "Night", ]

# T-test for Severity between Day and Night
t_test_result <- t.test(day_accidents$Severity, night_accidents$Severity)
t_test_result
## 
##  Welch Two Sample t-test
## 
## data:  day_accidents$Severity and night_accidents$Severity
## t = -25.029, df = 4376029, p-value < 2.2e-16
## alternative hypothesis: true difference in means is not equal to 0
## 95 percent confidence interval:
##  -0.010430544 -0.008915583
## sample estimates:
## mean of x mean of y 
##  2.209396  2.219069
# Base R boxplot
boxplot(day_accidents$Severity, night_accidents$Severity,
        names = c("Day", "Night"),
        main = "Accident Severity: Day vs Night",
        ylab = "Severity",
        col = c("lightblue", "salmon"),
        border = "darkblue")