# load packages
library(countdown)
library(tidyverse)
library(janitor)
library(colorspace)
library(broom)
library(fs)
# set theme for ggplot2
ggplot2::theme_set(ggplot2::theme_minimal(base_size = 14))
# set figure parameters for knitr
knitr::opts_chunk$set(
fig.width = 7, # 7" width
fig.asp = 0.618, # the golden ratio
fig.retina = 3, # dpi multiplier for displaying HTML output on retina
fig.align = "center", # center align figures
dpi = 300 # higher dpi, sharper image
)Visualizing time series data - II
Lecture 6
Air Quality Index
The AQI is the Environmental Protection Agency’s index for reporting air quality
Higher values of AQI indicate worse air quality
First look
This plot looks quite bizarre. What might be going on?
Another look
Improve
How would you improve this visualization?
Visualizing Durham AQI - v1
Recreate the following visualization.
Visualizing Durham AQI - v2
dch_2022 |>
ggplot(aes(x = date, y = aqi_value, group = 1)) +
geom_line(linewidth = 0.8) +
scale_x_date(
name = NULL, date_labels = "%b",
limits = c(ymd("2022-01-01"), ymd("2023-03-01"))
) +
scale_y_continuous(breaks = c(0, 50, 100, 150, 200, 300, 400)) +
geom_text(
data = aqi_levels,
aes(
x = ymd("2023-02-28"), y = aqi_mid,
label = level, color = darken(color, 0.3)
),
hjust = 1, size = 6, fontface = "bold"
) +
scale_color_identity() +
annotate(
geom = "text",
x = c(ymd("2022-01-01"), ymd("2023-01-01")), y = -80,
label = c("2022", "2023"), size = 4
) +
coord_cartesian(clip = "off", ylim = c(0, 400)) +
labs(
x = NULL, y = "AQI",
title = "Ozone and PM2.5 Daily AQI Values",
subtitle = "Durham-Chapel Hill, NC",
caption = "\nSource: EPA Daily Air Quality Tracker"
) +
theme(
plot.title.position = "plot",
panel.grid.minor.y = element_blank(),
panel.grid.minor.x = element_blank()
)What’s missing?
What’s missing in this plot?
Filling in the gaps
Plotting cumulatives
dch_2022 |>
select(date, aqi_value) |>
filter(!is.na(aqi_value)) |>
arrange(date) |>
mutate(
good_aqi = if_else(aqi_value <= 50, 1, 0),
cumsum_good_aqi = cumsum(good_aqi)
) |>
ggplot(aes(x = date, y = cumsum_good_aqi, group = 1)) +
geom_line() +
scale_x_date(date_labels = "%b %Y") +
labs(
x = NULL, y = "Number of days",
title = "Cumulative number of good AQI days (AQI < 50)",
subtitle = "Durham-Chapel Hill, NC",
caption = "\nSource: EPA Daily Air Quality Tracker"
) +
theme(plot.title.position = "plot")Plot trend since 2016
dch |>
ggplot(aes(x = date, y = cumsum_good_aqi, group = 1)) +
geom_smooth(method = "lm", color = "pink") +
geom_line() +
scale_x_date(
expand = expansion(mult = 0.07),
date_labels = "%Y"
) +
labs(
x = NULL, y = "Number of days",
title = "Cumulative number of good AQI days (AQI < 50)",
subtitle = "Durham-Chapel Hill, NC",
caption = "\nSource: EPA Daily Air Quality Tracker"
) +
theme(plot.title.position = "plot")`geom_smooth()` using formula = 'y ~ x'Visualize detrended data
dch_aug |>
ggplot(aes(x = date, y = ratio, group = 1)) +
geom_hline(yintercept = 1, color = "gray") +
geom_line() +
scale_x_date(
expand = expansion(mult = 0.07),
date_labels = "%Y"
) +
labs(
x = NULL, y = "Number of days\n(detrended)",
title = "Cumulative number of good AQI days (AQI < 50)",
subtitle = "Durham-Chapel Hill, NC",
caption = "\nSource: EPA Daily Air Quality Tracker"
) +
theme(plot.title.position = "plot")Plot trend since 2016
sf |>
ggplot(aes(x = date, y = cumsum_good_aqi, group = 1)) +
geom_smooth(method = "lm", color = "pink") +
geom_line() +
scale_x_date(
expand = expansion(mult = 0.07),
date_labels = "%Y"
) +
labs(
x = NULL, y = "Number of days",
title = "Cumulative number of good AQI days (AQI < 50)",
subtitle = "San Francisco-Oakland-Hayward, CA",
caption = "\nSource: EPA Daily Air Quality Tracker"
) +
theme(plot.title.position = "plot")`geom_smooth()` using formula = 'y ~ x'Visualize detrended data
sf_aug |>
ggplot(aes(x = date, y = ratio, group = 1)) +
geom_hline(yintercept = 1, color = "gray") +
geom_line() +
scale_x_date(
expand = expansion(mult = 0.07),
date_labels = "%Y"
) +
labs(
x = NULL, y = "Number of days\n(detrended)",
title = "Cumulative number of good AQI days (AQI < 50)",
subtitle = "San Francisco-Oakland-Hayward, CA",
caption = "\nSource: EPA Daily Air Quality Tracker"
) +
theme(plot.title.position = "plot")Detrending
In step 2 we fit a very simple model
Depending on the complexity you’re trying to capture you might choose to fit a much more complex model
You can also decompose the trend into multiple trends, e.g. monthly, long-term, seasonal, etc.
