Add formatting to numbers output by bi_class_breaks()

[RickPack]

Might you enable the passing of functions to bi_class_breaks() to format numbers?

Or, for my particular interest, you could add an SI_lab (International System of Units) parameter that would pass digits to the si_number function in my code below.

This reduction of text for large numbers makes it easier to increase the font size and position the legend.

Currently

With proposed SI_lab parameter

Using SI_lab(3)

Code to aid addition of SI_lab parameter

First, create miami_rent_income object using my code at https://stackoverflow.com/a/72591940/2305061.

Then, focusing on modification to labels1.

miami_rent_income_bivar <-
  biscale::bi_class(miami_rent_income,
                    x = Median_Rent,
                    y = Median_Income, 
                    style = "jenks",
                    dim = 3)

labels1 <- bi_class_breaks(miami_rent_income, 
                           x = Median_Rent, 
                           y = Median_Income, 
                           style = "jenks", 
                           dim = 3, dig_lab = c(20,20), split = FALSE)

## function found at https://stackoverflow.com/a/59086755/2305061
si_number = function(x, digits) {
  
  compress = function(x, n) {
    signif(x * 10^(-n), digits)
  }
  
  case_when(
    ## decimals like 0.33
    abs(x) <    1   ~ as.character(x),   
    abs(x) >= 1e6   ~ paste0(compress(x, 6), "M"),
    abs(x) >= 1000  ~ paste0(compress(x, 3), "k"),
    abs(x) >= 1     ~ as.character(compress(x, 0)),
    abs(x) >= 0.001 ~ paste0(compress(x, -3), "m"),
    abs(x) >= 1e-6  ~ paste0(compress(x, -6), "u"),
    ## number of 0
    TRUE ~ as.character(x)
  )
}

labels1$bi_x <- paste0(si_number(as.numeric(str_extract(labels1$bi_x, ".+?(?=-)")), 3)
                       ,"-",
                       si_number(as.numeric(str_extract(labels1$bi_x, "(?<=\\-).*")), 3)
)
labels1$bi_y <- paste0(si_number(as.numeric(str_extract(labels1$bi_y, ".+?(?=-)")), 3)
                       ,"-",
                       si_number(as.numeric(str_extract(labels1$bi_y, "(?<=\\-).*")), 3)
)


miami_rent_income_bivar_legend <- bi_legend(pal = "DkCyan",
                                             breaks = labels1,
                                             xlab = "Median Rent ($)",
                                             ylab = "Median Income ($)",
                                             size = 08)

[RickPack]

If you like my idea, I offer to attempt the addition of this SI_lab parameter to bi_class_breaks()

[chris-prener]

thanks for the suggestion @RickPack! I'd love to add a formatting argument for bi_class_braks(). I had thought about it for the v1.0 release last month but ran out of steam, so this is most welcome. Can you mirror the dig_lab functionality so that it takes either a scalar that is applied to both x and y or a vector where length(vector) == 2?

[RickPack]

@chris-prener, I'll give it a try in the next few days!

[RickPack]

Can you assign this to me?

[RickPack]

I will get to this soon, @chris-prener, hopefully later today.

[chris-prener]

No rush @RickPack!

[RickPack]

Working on this now!

[RickPack]

@chris-prener, can you help me understand the need for as.numeric under the split == TRUE condition in https://github.com/chris-prener/biscale/blob/main/R/bi_class.R? My use of the SI prefix is causing an error (e.g., 323k instead of 323001).

[RickPack]

@chris-prener also noting your title at https://chris-prener.github.io/biscale/articles/biscale.html has "Bivarite"

[RickPack]

Along with the "Bivarite" misspelling, also need a library(ggplot2) somewhere before the ggplot() at https://chris-prener.github.io/biscale/articles/biscale.html.

[chris-prener]

I'm going to leave this open just so that I can iron out some documentation for how this works, @RickPack - will ask you to take a look at that when it is done!

[RickPack]

@chris-prener, will do and unfortunately, my technique occasionally does not work properly. I am on the run but will get you an example and hopefully a solution as soon as I can. It generally works though.

[RickPack]

Example of a problem. Notice the "300m" label on the x-axis.

[quintercept]

Hi there! Been following this thread closely and excited you've been working on some formatting for the bi_class_breaks() labels. I wanted to see if there is any update or if this has been added to the newest version or if there's plans to released it yet? Thanks so much!

[chris-prener]

Hi @quintercept this is still in process!

[RickPack]

@quintercept , you or another can edit my pull request if you like? I am not sure when I will get to this. Along with fixing the issue above, I would like to experiment with either label_number(scale_cut = cut_short_scale()) or label_number(scale_cut = cut_si("unit")) from the scales package.

[RickPack]

Can anyone reproduce the problem I presented using the addition of an SI_lab parameter to bi_class_breaks() in my pull request? I have had a hard time doing so. I think it was for a year from 2018 - 2021, Census ACS5 data (default for the tidyverse R package), Fairfield County, Connecticut.