Making population pyramids with Plotly

[walkerke]

First and foremost - thanks for the great work on the package, the API is fantastic!

I commonly build population pyramids and I'd like to use Plotly to do this as well. However, I'm running into a couple issues when doing this, so I thought I'd see if anyone had some feedback on these issues.

Normally when setting up a population pyramid I set the population values for one of the genders to negative numbers to built a "pyramid" around a central axis. For example:

library(plotly)

age <- rep(1:5, 2)

sex <- rep(c('Male', 'Female'), each = 5)

pop <- c(-1000, -944, -888, -762, -667, 1100, 999, 844, 789, 655)

df <- data.frame(age, sex, pop)

plot_ly(df, x = pop, y = age, group = sex, type = 'bar', orientation = 'h')

On the right track here, but I want the bars to align, as in a stacked bar chart. When I change the barmode to 'stack', I get this:

plot_ly(df, x = pop, y = age, group = sex, type = 'bar', orientation = 'h') %>%
    layout(barmode = 'stack')

Alternatively, I also tried adding the genders as separate traces, but Plotly then plots positive numbers as negative:

library(dplyr)

male <- filter(df, sex == 'Male')

female <- filter(df, sex == 'Female')

plot_ly(male, x = pop, y = age, type = 'bar', orientation = 'h') %>%
    add_trace(female, type = 'bar', x = pop, y = age, orientation = 'h') %>%
    layout(barmode = 'stack')

If you have any ideas about this, please let me know!

Also: when creating population pyramids, I want to display all of the axis values as absolute values on either side of zero. In other R/JS libraries, I've passed something like the following to tickformat:

function(d) { return d3.format(',.1f')(Math.abs(d)) }

I've tried this out in Plotly, both as a string and wrapped in htmlwidgets::JS(), like this:

plot_ly(df, x = pop, y = age, group = sex, type = 'bar', orientation = 'h') %>%
  layout(xaxis = list(
    tickformat = htmlwidgets::JS("function(d) { return d3.format(',.0f')(Math.abs(d)) }")
  ))

and I don't get it to change. I'll keep working on this, but please let me know if you have any pointers!

[cpsievert]

This will get you most of the way there:

plot_ly(df, x = pop, y = age, group = sex, type = 'bar', orientation = 'h') %>% 
  layout(bargap = 0, barmode = "overlay")

I'm not sure if it's possible to display absolute values in the tick labels. Passing a JS function is an interesting idea, but I don't think plotly.js supports that.

For future reference, we'd like to use issues for just bug reports and feature requests -- https://github.com/ropensci/plotly/blob/master/CONTRIBUTING.md

Feel free to request absolute value tick labels here

[etpinard]

@walkerke you can use tickvals and ticktext to fully customize your graphs's tick values.

[walkerke]

Excellent, thank you for the quick reply! I'll try this out.

[walkerke]

Tried it out and everything is working great! Fantastic library to work with. For anyone who might come across this issue and is looking for the solution:

library(plotly)
library(dplyr)

age <- rep(1:5, 2)

sex <- rep(c('Male', 'Female'), each = 5)

pop <- c(-1000, -944, -888, -762, -667, 1100, 999, 844, 789, 655)

df <- data.frame(age, sex, pop) %>%
  mutate(abs_pop = abs(pop))

plot_ly(df, x = pop, y = age, group = sex, type = 'bar', orientation = 'h',
        hoverinfo = 'y+text+name', text = abs_pop) %>%
  layout(bargap = 0.1, barmode = 'overlay',
         xaxis = list(tickmode = 'array', tickvals = c(-1000, -500, 0, 500, 1000),
                      ticktext = c('1000', '500', '0', '500', '1000')))

[jgonzale]

@walkerke, I've tried to reproduce this last plot you presented, but I got stuck with this error:

Warning in plot_ly(df, x = ~pop, y = ~age, group = ~sex, type = "bar", orientation = "h", : The group argument has been deprecated. Use `group_by()` or split instead.

I'm not an R expert, and I don't know how to proceed.

Many thanks!

[pssguy]

@jgonzale
This code will not work with latest version of plotly. There is a lot of coverage of plotly for R including a webinar, book, many examples and a reference guide. All worth investing time on. This should get the basic chart up for you

library(plotly)
library(dplyr)

age <- rep(1:5, 2)

sex <- rep(c('Male', 'Female'), each = 5)

pop <- c(-1000, -944, -888, -762, -667, 1100, 999, 844, 789, 655)

df <- data.frame(age, sex, pop) %>%
  mutate(abs_pop = abs(pop))


df %>% 
  plot_ly(x= ~pop, y=~age,color=~sex) %>% 
  add_bars(orientation = 'h', hoverinfo = 'text', text = ~abs_pop) %>%
  layout(bargap = 0.1, barmode = 'overlay',
         xaxis = list(tickmode = 'array', tickvals = c(-1000, -500, 0, 500, 1000),
                      ticktext = c('1000', '500', '0', '500', '1000')))

[jgonzale]

@pssguy, thank you very much!!

I'm so fortunate there is a large chunk of documentation about plotly out there! (And it's integration with R Markdown etc. etc.)

Great!