University of Iowa

College of Liberal Arts and Sciences

• STAT:4580
• Syllabus
• Outline
• Homework
• Notes
• Resources
-->

Assignment 4 Notes

General Issues

• Make sure you name your files as requested, including matching the specified use of upper and lower case. This matters on file systems that are case-sensitive.

• Make sure to commit your work to your local repository and push your commits to GitLab. We can only see what is on GitLab, not what is on your computer. You can check what we see by going to the GitLab web interface.

• Include your name and the date in the header of your .Rmd file using author: and date: tags.

• Your HTML file should be a report of your findings.

  • Any graph you show should be discussed in your narrative.

  • Any code you show should be discussed in your narrative.

  • If you do not need to discuss a piece of code in the narrative, use echo FALSE to avoid showing it.

• If you load a file that you have included in your repository or that you download to your repository then you need to make sure the code in your Rmarkdown document uses a relative path, not an absolute one. Absolute paths will only make sense on your computer, not on the computer of someone else who downloads your repository.

• If you want to check your work is reproducible you can download your work to a computer other than the one you use for developing it. One option is the CLAS Linux systems accessed via FastX. You can use RStudio there to set up a clean copy of your repository and then just pull your changes and check that they knit successfully. Using STAT4580::checkHW is a convenient way to do this.

1. Evaluate a Visualization

The Vox visualization attracted some attention in the internet; some examples:

• A post on https://www.iflscience.com/.

• A post on https://nonprofitquarterly.org with a link to one alternative visualization.

• Another alternative visualization.

• The original visualization made the very common mistake of mapping magnitudes to circle radius, which distorts the perceived magnitudes since perception focuses on area. The main change in the revision is to map magnitude to area.

• The revision also changed some color assignments, but kept the traditional assignment of pink for Breast Cancer.

Analysis of the visualization:

• Items: diseases and associated measurements.

• Attributes: disease, money raised; deaths.

• Marks: circles, text.

• Channels: vertical position, area, color (hue), text.

• Mappings:

  • Ranks within the numeric variables are mapped to vertical position.

  • Magnitudes of numeric variables are mapped to circle areas.

  • Magnitudes are also mapped to text labels.

  • Disease is mapped to color (hue).

A goal of the visualization is to show the discrepancy between the relative amounts raised and the relative numbers of deaths. This relation is communicated by matching the positions or sizes of the corresponding circles by color, a weaker channel.

One good alternative, used in one of the links above, is a scatter plot:

Other options:

• a Tufte-style slope graph using standardized variables or ranks (essentially a parallel coordinates plot; used in another of the links above);

• visualizing a derived variable, such as funds per death.

There are issues with the data; some of these are discussed in the articles linked to above.

2. EPA Fuel Economy Data

library(lubridate)
library(readr)
if (! file.exists("vehicles.csv.zip") ||
    file.mtime("vehicles.csv.zip") + months(6) < now())
    download.file("http://www.stat.uiowa.edu/~luke/data/vehicles.csv.zip",
                  "vehicles.csv.zip")
newmpg <- read_csv("vehicles.csv.zip", guess_max = 100000)

From the documentation for the data the appropriate variables seem to be:

• fuelType1 represents the primary fuel type, fl in mpg.
• highway08 corresponds to hwy in mpg;
• cylinders corresponds to cyl in mpg;
• displ corresponds to displ in mpg;

The primary fuel type counts are

library(dplyr)
tbl <- count(newmpg, fuelType1)
kbl <- knitr::kable(tbl, format = "html")
kableExtra::kable_styling(kbl, full_width = FALSE)

fuelType1	n
Diesel	1274
Electricity	649
Midgrade Gasoline	162
Natural Gas	60
Premium Gasoline	14655
Regular Gasoline	30275

A bar chart of these numbers:

thm <- theme_minimal() + theme(text = element_text(size = 16))
ggplot(tbl, aes(x = n, y = reorder(fuelType1, n))) +
    geom_col() +
    scale_x_continuous(expand = expansion(mult = c(0, .1))) +
    thm +
    ylab(NULL)

Regular gas is the dominant fuel type over all years, with premium second. All other fuel types, including electricity, make up a small fraction.

3. Fuel Type Over the Years

A filled bar chart shows changes in the primary fuel type used over the years:

newmpg2 <- filter(newmpg, year <= 2023) |>
    mutate(year = factor(year))
ggplot(newmpg2, aes(y = year, fill = fuelType1)) +
    geom_bar(position = "fill") +
    scale_x_continuous(expand = c(0, 0)) +
    labs(x = "Proportion", y = NULL)

Regular gas was the predominant fuel type in the mid 1980s, but premium’s share has gradually increased to the point where almost as many models use premium as regular. Diesel’s popularity declined early and had a small resurgence recently. The market share for electricity is still quite small but is growing.

4. Highway Fuel Economy Over the Years

newmpg3 <- filter(newmpg, year <= 2023, year >= 2000) |>
    mutate(year = factor(year))
alpha <- 0.2
size <- 0.3
nyear <- length(levels(newmpg3$year))

A strip chart is a useful way to look at the full data for a numeric variable at several different levels of a discrete variable, but some tuning is needed for larger data sets. For examining 24 years of highway gas mileage data from the EPA data set using alpha = 0.2 and size = 0.3 along with jittering seems to work reasonably well:

ggplot(newmpg3, aes(x = highway08, y = year)) +
    geom_point(position = "jitter", size = size, alpha = alpha) +
    ylab(NULL) +
    thm

Over time the highway gas mileage distributions are moving upward a little bit, with the upper tails becoming gradually longer and an increasing number of very high efficiency models (mostly electric).

This work is licensed under the CC BY-NC 3.0 Creative Commons License.