In mikeroswell/ratioScales: Rational Scales for Visualizing Change

knitr::opts_chunk$set(
  collapse = TRUE,
  comment = "#>"
  , fig.width = 7
  , fig.height = 7
)

library(ratioScales)
library(dplyr)
library(tidyr)
theme_set(theme_classic() +
          theme(axis.text.x = element_text(angle = 45,
                                           vjust = 1, hjust = 1)))

Hook to illustrate a problem that needs to be solved (perhaps not exchange-rate focused)

Introducing the divMult scale

# first, without logarithmic rescaling

# second, with log10 transformation of y-axis

``` {r divMult plot}

<!-- Then, want to provide percDiff as a good compromise as well -->

# Percent differences can work too, if an appropriate scale is used

The problem with percent differences (as typically presented) is that they suggest
symmetry in confusing ways, and make the large relative changes near 0 and 100 
seem small. But, if an appropriate (i.e., logarithmic) scaling is used, familiar
percent differences make great axis ticks. 



```r

mikeroswell/ratioScales documentation built on July 18, 2024, 6:36 p.m.