Class 17

Projects and More Visualization

Materials for class on

2024-10-24

Agenda

Today we’ll focus on:

hp1 work/troubleshooting time
more dataviz

Homework project 1 notes

make sure to be clear what you are trying to do/show with in text
use functions in a way that makes sense for your analysis, or “does something” - don’t just show that you can randomly do a join, do something with it

Scale Transformations

How would you represent something on a log scale, rather than a linear scale? For reference, the log(arithmic) scale is a way of transforming numbers that have equal spacing between exponents, such as 10, 100, 1000, etc. This is used for a variety of purposes, including the Richter magnitude scale (for earthquakes).

log(10, base = 10)

#> [1] 1

log(100, base = 10)

#> [1] 2

log(1000, base = 10)

#> [1] 3

Some data show an exponential pattern:

# create some exponential data 
set.seed(343)
df_exp <- tibble(x = letters, y = sort(rexp(26, rate = .1)))
ggplot(df_exp, aes(x = x, y = y, group = 1)) +
  geom_point() +
  geom_smooth(method = loess)

It doesn’t map well to a linear analysis:

ggplot(df_exp, aes(x = x, y = y, group = 1)) +
  geom_point() +
  geom_smooth(method = lm)

It can be useful to transform the values to a log scale. Depending on the purpose, this can be done by mutating the data, or by changing the axis scale:

ggplot(df_exp, aes(x = x, y = log(y), group = 1)) +
  geom_point() +
  geom_smooth(method = lm)

Using scale_y_log10(), the values are shown on the linear scale, but spaced according to the log scale:

ggplot(df_exp, aes(x = x, y = y, group = 1)) +
  geom_point() +
  scale_y_log10() +
  geom_smooth(method = lm)

There are many other transformations you can do with scales in ggplot2 depending on your use case, including custom ones. The reference is in the scales documentation. Here’s an example using a different log scale:

ggplot(df_exp, aes(x = x, y = y, group = 1)) +
  geom_point() +
  scale_y_continuous(transform = scales::log2_trans()) +
  geom_smooth(method = lm)

You can also change the coordinates, which is done after statistics. There is a good demonstration of these differences in the documentation of coord_trans.

Exploring bikes data

In order to setup the “bikes” data for the materials, use this code:

bikes <- readr::read_csv("https://raw.githubusercontent.com/z3tt/graphic-design-ggplot2/main/data/london-bikes-custom.csv",
  col_types = "Dcfffilllddddc"
)

bikes$season <- forcats::fct_inorder(bikes$season)

Example plot from the tutorial:

  ggplot(
    bikes,
    aes(x = temp_feel, y = count,
        color = season,
        group = day_night)
  ) +
  geom_point(
    alpha = .5
  ) +
  geom_smooth(
    method = "lm",
    color = "black"
  ) +
  labs(
    x = "Feels-like temperature (°F)",
    y = "Reported bike shares",
    title = "TfL bike sharing trends",
    color = "Season:"
  ) +
  theme(
  panel.grid.minor = element_blank(),
  plot.title = element_text(face = "bold"),
  legend.position = "top",
  plot.title.position = "plot"
)

Post-class material

We will use Cédric Scherer’s slides and work through those examples live in class. I will post any additional code from class here afterwards.

--- title: "Class 17" subtitle: "Projects and More Visualization" date: 2024-10-24 date-format: "YYYY-MM-DD" editor: markdown: wrap: 72 editor_options: chunk_output_type: console execute: message: false warning: false --- ```{r} #| echo: FALSE #| include: false library(tidyverse) ``` ## Further Reading No preparation reading was assigned, but we will/may refer to this material today: - {{< fa external-link >}} [Cédric Scherer's Concepts of ggplot2, pt 1](https://rstudio-conf-2022.github.io/ggplot2-graphic-design/modules/d1-02_concepts_pt1.html) ## Agenda Today we'll focus on: - hp1 work/troubleshooting time - more dataviz ## Homework project 1 notes - make sure to be clear what you are trying to do/show with in text - use functions in a way that makes sense for your analysis, or "does something" - don't just show that you can randomly do a join, do something with it ## Scale Transformations How would you represent something on a log scale, rather than a linear scale? For reference, the log(arithmic) scale is a way of transforming numbers that have equal spacing between exponents, such as 10, 100, 1000, etc. This is used for a variety of purposes, including the Richter magnitude scale (for earthquakes). ```{r} log(10, base = 10) log(100, base = 10) log(1000, base = 10) ``` Some data show an exponential pattern: ```{r} # create some exponential data set.seed(343) df_exp <- tibble(x = letters, y = sort(rexp(26, rate = .1))) ggplot(df_exp, aes(x = x, y = y, group = 1)) + geom_point() + geom_smooth(method = loess) ``` It doesn't map well to a linear analysis: ```{r} ggplot(df_exp, aes(x = x, y = y, group = 1)) + geom_point() + geom_smooth(method = lm) ``` It can be useful to transform the values to a log scale. Depending on the purpose, this can be done by mutating the data, or by changing the axis scale: ```{r} ggplot(df_exp, aes(x = x, y = log(y), group = 1)) + geom_point() + geom_smooth(method = lm) ``` Using `scale_y_log10()`, the values are shown on the linear scale, but spaced according to the log scale: ```{r} ggplot(df_exp, aes(x = x, y = y, group = 1)) + geom_point() + scale_y_log10() + geom_smooth(method = lm) ``` There are many other transformations you can do with scales in ggplot2 depending on your use case, including custom ones. The reference is in the [scales documentation](https://scales.r-lib.org/reference/index.html#transformations). Here's an example using a different log scale: ```{r} ggplot(df_exp, aes(x = x, y = y, group = 1)) + geom_point() + scale_y_continuous(transform = scales::log2_trans()) + geom_smooth(method = lm) ``` You can also change the coordinates, which is done after statistics. There is a good demonstration of these differences in the [documentation of coord_trans](https://ggplot2.tidyverse.org/reference/coord_trans.html). ## Exploring bikes data In order to setup the "bikes" data for the materials, use this code: ```{r} bikes <- readr::read_csv("https://raw.githubusercontent.com/z3tt/graphic-design-ggplot2/main/data/london-bikes-custom.csv", col_types = "Dcfffilllddddc" ) bikes$season <- forcats::fct_inorder(bikes$season) ``` Example plot from the tutorial: ```{r} ggplot( bikes, aes(x = temp_feel, y = count, color = season, group = day_night) ) + geom_point( alpha = .5 ) + geom_smooth( method = "lm", color = "black" ) + labs( x = "Feels-like temperature (°F)", y = "Reported bike shares", title = "TfL bike sharing trends", color = "Season:" ) + theme( panel.grid.minor = element_blank(), plot.title = element_text(face = "bold"), legend.position = "top", plot.title.position = "plot" ) ``` ## Post-class material We will use Cédric Scherer's slides and work through those examples live in class. I will post any additional code from class here afterwards.