Objective:
You will use R to analyze the built-in airquality dataset, applying descriptive statistics techniques to explore environmental data. The assignment covers measures of central tendency, spread, histograms, boxplots, scatterplots, correlations, and summary tables, aligning with the Week 6 agenda on Descriptive Statistics.
Dataset
Source: Built-in R dataset airquality.
Description: Contains 153 observations of daily air quality measurements in New York from May to September 1973.
Variables (selected for this assignment):
Notes
-The airquality dataset has missing values in Ozone and Solar.R. The code uses na.rm = TRUE or use = “complete.obs” to handle them.
-If you encounter errors, check that tidyverse and corrplot are installed and loaded.
-Feel free to modify plot aesthetics (e.g., colors, binwidth) to enhance clarity.
Instructions:
Complete the following tasks using R to analyze the airquality dataset. Submit your Rpubs link that includes code, outputs (tables and plots), and written interpretations for each task. Ensure you load the dataset using data(airquality) and install/load the tidyverse and corrplot packages.
#Load your dataset
library(tidyverse)
## ── Attaching core tidyverse packages ──────────────────────── tidyverse 2.0.0 ──
## ✔ dplyr 1.1.4 ✔ readr 2.1.5
## ✔ forcats 1.0.1 ✔ stringr 1.5.2
## ✔ ggplot2 4.0.0 ✔ tibble 3.3.0
## ✔ lubridate 1.9.4 ✔ tidyr 1.3.1
## ✔ purrr 1.1.0
## ── 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 errors
library(corrplot)
## corrplot 0.95 loaded
data("airquality")
Tasks and Questions
Using functions you learned this week, compute mean, median, standard deviation, min, and max separately for Ozone, Temp, and Wind.
colSums(is.na(airquality))
## Ozone Solar.R Wind Temp Month Day
## 37 7 0 0 0 0
#airquality <- airquality %>%
#mutate(Ozone_imputed = if_else(
#is.na(Ozone), mean(Ozone, na.rm = TRUE), Ozone))
summary_stats_ozone <- airquality |>
summarise(
Mean_Ozone = mean(airquality$Ozone, na.rm = TRUE),
Median_Ozone = median(airquality$Ozone, na.rm = TRUE),
SD_Ozone = sd(airquality$Ozone, na.rm = TRUE),
Min_Ozone = min(airquality$Ozone, na.rm = TRUE),
Max_Ozone = max(airquality$Ozone, na.rm = TRUE))
summary_stats_ozone
## Mean_Ozone Median_Ozone SD_Ozone Min_Ozone Max_Ozone
## 1 42.12931 31.5 32.98788 1 168
summary_stats_temp <- airquality |>
summarise(
Mean_Temp = mean(airquality$Temp, na.rm = TRUE),
Median_Temp = median(airquality$Temp, na.rm = TRUE),
SD_Temp = sd(airquality$Temp, na.rm = TRUE),
Min_Temp = min(airquality$Temp, na.rm = TRUE),
Max_Temp = max(airquality$Temp, na.rm = TRUE))
summary_stats_temp
## Mean_Temp Median_Temp SD_Temp Min_Temp Max_Temp
## 1 77.88235 79 9.46527 56 97
summary_stats_wind<- airquality |>
summarise(
Mean_Wind = mean(airquality$Wind, na.rm = TRUE),
Median_Wind = median(airquality$Wind, na.rm = TRUE),
SD_Wind = sd(airquality$Wind, na.rm = TRUE),
Min_Wind = min(airquality$Wind, na.rm = TRUE),
Max_Wind = max(airquality$Wind, na.rm = TRUE))
summary_stats_wind
## Mean_Wind Median_Wind SD_Wind Min_Wind Max_Wind
## 1 9.957516 9.7 3.523001 1.7 20.7
Question: Compare the mean and median for each
variable. Are they similar or different, and what does this suggest
about the distribution (e.g., skewness)? What does the standard
deviation indicate about variability?
The mean is greater than
the median for the ozone variable, indicating a right-skewed
distribution. The mean is a bit less than the median for the temp
variable, indicating a slight left-skew. The wind variable’s mean is
slightly greater than the median, however it is close enough to assume
that this is a mainly symmetrical distribution that is not heavily
skewed one way or the other.
The standard deviation for the ozone variable is ~33 indicating that
the data is widely scattered around the mean. The temp variable has a
standard deviation of ~9.5, indicating a fair amount of spread but less
than the ozone variable. The wind variable has a standard deviation of
~3.5, indicating a fairly low spread of data around the mean.
Generate the histogram for Ozone.
hist(airquality$Ozone, main = "Histogram of Ozone Levels",
xlab = "Value", col = "lightblue", breaks = 20)
Question: Describe the shape of the ozone
distribution (e.g., normal, skewed, unimodal). Are there any outliers or
unusual features?
The distribution is right-skewed, with most
of the ozone values clustered below 50 and a long tail extending toward
the higher values. It is unimodal with it’s peak around 10-20. There are
some outliers above 150.
Create a boxplot of ozone levels (Ozone) by month, with months
displayed as names (May, June, July, August, September) instead of
numbers (5–9). Recode the Month variable into a new column called
month_name with month names using case_when
from week 4.Generate a boxplot of Ozone by month_name.
airquality <- airquality %>%
mutate(month_name = case_when(
Month == 5 ~ "May",
Month == 6 ~ "June",
Month == 7 ~ "July",
Month == 8 ~ "August",
Month == 9 ~ "September"
))
airquality$month_name <- factor(airquality$month_name, levels = c("May", "June", "July", "August", "September"))
ggplot(airquality, aes(x = month_name, y = Ozone)) +
geom_boxplot(fill = "skyblue", color = "black") +
labs(title = "Ozone Levels by Month",
x = "Month", y = "Ozone Levels") +
theme_minimal()
## Warning: Removed 37 rows containing non-finite outside the scale range
## (`stat_boxplot()`).
Question: How do ozone levels vary across months?
Which month has the highest median ozone? Are there outliers in any
month, and what might they indicate?
The ozone levels vary
widely across the months, with it peaking during the summer in July and
August while May, June, and September show lower levels. July has the
highest median levels, showing that July experienced the highest typical
ozone concentrations. There are noticeable outliers in May, June,
August, and September; with September having the greatest amount of
outliers. These outliers might indicate short, transient events which
spiked ozone levels.
Produce the scatterplot of Temp vs. Ozone, colored by Month.
ggplot(airquality, aes(x = Temp, y = Ozone, color = factor(month_name))) +
geom_point(size = 3) +
scale_color_manual(values = c(
"May" = "red",
"June" = "orange",
"July" = "yellow",
"August" = "green",
"September" = "blue"
), labels = c("May", "June", "July", "August", "September")) +
labs(title = "Ozone Levels vs. Monthly Temperature",
x = "Temperature (F)", y = "Ozone",
color = "Month") +
theme_minimal()
## Warning: Removed 37 rows containing missing values or values outside the scale range
## (`geom_point()`).
Question: Is there a visible relationship between
temperature and ozone levels? Do certain months cluster together (e.g.,
higher ozone in warmer months)? Describe any patterns.
There
is a noticeable positive relationship between temperature and ozone
levels. As we can see in the graph, excluding the extreme outliers, the
further right the points are the higher up they tend to be. We can also
see that there is some evidence of clustering, as seen with May and June
having generally lower ozone values and cooler temperatures, July and
August showing higher ozone levels and warmer temperatures, and
September shifting back toward lower temperatures and lower ozone
levels. The presence of outliers could indicate that there are other
conditions that can cause a transient burst of ozone levels.
Compute and visualize the correlation matrix for Ozone, Temp, and Wind.
cor_matrix <- cor(
airquality |>
select(Ozone, Temp, Wind), use = "complete.obs")
cor_matrix
## Ozone Temp Wind
## Ozone 1.0000000 0.6983603 -0.6015465
## Temp 0.6983603 1.0000000 -0.5110750
## Wind -0.6015465 -0.5110750 1.0000000
corrplot(cor_matrix, method = "color", type = "upper", order = "hclust",
tl.col = "black", tl.srt = 45, addCoef.col = "black",
title = "Correlation Matrix of Numeric Variables")
Question: Identify the strongest and weakest
correlations. For example, is ozone more strongly correlated with
temperature or wind speed? Explain what the correlation values suggest
about relationships between variables.
Ozone and temperature
shows the strongest correlation with an r value of 0.70. This means that
as the temperature increases, the ozone levels do as well, as seen in
the previous scatterplot. The weakest correlation is between wind and
temperature, with an r value of -0.51. This suggest a modest negative
relationship, indicating that higher wind speeds are somewhat associated
with lower temperatures. Ozone and wind have a negative correlation as
well, with and r value of -0.60. This suggests that higher wind speeds
have the effect of lowering the ozone values.
Generate the summary table grouped by Month. Generate the summary table grouped by Month. It should include count, average mean of ozone, average mean of temperature, and average mean of wind per month.
summary_table <- airquality |>
group_by(month_name) |>
summarise(
Count = n(),
Avg_Ozone = mean(Ozone, na.rm = TRUE),
Avg_Temp = mean(Temp, na.rm = TRUE),
Avg_Wind = mean(Wind, na.rm = TRUE),
)
summary_table
## # A tibble: 5 × 5
## month_name Count Avg_Ozone Avg_Temp Avg_Wind
## <fct> <int> <dbl> <dbl> <dbl>
## 1 May 31 23.6 65.5 11.6
## 2 June 30 29.4 79.1 10.3
## 3 July 31 59.1 83.9 8.94
## 4 August 31 60.0 84.0 8.79
## 5 September 30 31.4 76.9 10.2
Question: Which month has the highest average ozone
level? How do temperature and wind speed vary across months? What
environmental factors might explain these differences?
August
has the highest average ozone level, sitting at ~60. The temperatures
increase from May through August, and then drop in September. Wind
speeds are highest in May, Juny, and September, with July and August
showing a decline. This shows us that higher temperatures and lower wind
speeds seem to have an effect in producing an overall greater amount of
ozone. Hot summer conditions seem to be exacerbating ozone levels
compared to cooler and more windy periods of the year.