sunburstR

I fell in love with this Sequences Sunburst by Kerry Rodden immediately when it wastweeted to the world in October 2013. I’ll put it in an iframe below to insure that you don’t miss it. It was also mentioned it back in the post Week 3 | More Network Layouts.

The additional request by Mark Riseley in this Github issue motivated me to attempt to squeeze this beauty into an htmlwidget

Breaking My Rules

This week I broke a number of my rules or best practices for htmlwidgets. My excuses follow:

1. I really need help, input, feedback on this to determine direction and functionality. My twitter poll provided four potential use cases: hierarchical topic models from Carson Sievert, TraMineR state sequences from James Curley,thoroughbred sire pedigree family trees by Tom H, and VERIS security data by Bob Rudis.

2. I did two widgets this week, since I missed last week due to vacation, and I just ran out of time.

Quick Installation

sunburstR is not yet on CRAN and won’t be unless someone asks, so for now please install with devtools::install_github.

devtools::install_github("timelyportfolio/sunburstR")

Examples

Original Example but in R

The most obvious example is to recreate the original example but in R. Since I embedded the original in an iframe above, I’ll just post the code to demonstrate.

# devtools::install_github("timelyportfolio/sunburstR")

library(sunburstR)

# read in sample visit-sequences.csv data provided in source
#   https://gist.github.com/kerryrodden/7090426#file-visit-sequences-csv
sequence_data <- read.csv(
  paste0(
    "https://gist.githubusercontent.com/kerryrodden/7090426/"
    ,"raw/ad00fcf422541f19b70af5a8a4c5e1460254e6be/visit-sequences.csv"
  )
  ,header=F
  ,stringsAsFactors = FALSE
)

sunburst(sequence_data)

Example with TraMineR

TraMineR is a really, really nice R package for working with sequences.

Gabadinho, A., Ritschard, G., Müller, N.S. & Studer, M. (2011), Analyzing and visualizing state sequences in R with TraMineR, Journal of Statistical Software. Vol. 40(4), pp. 1-37.

Let’s adapt the first example from the TraMineR vignette so that we can visualize it with our new sunburstR. I apologize to the non-piping R world, but pipes make this so much easier.

library(TraMineR)
library(sunburstR)
library(pipeR)

# use example from TraMineR vignette
data("mvad")
mvad.alphab <- c(
  "employment", "FE", "HE", "joblessness",
  "school", "training"
)
mvad.seq <- seqdef(mvad, 17:86, xtstep = 6, alphabet = mvad.alphab)

# to make this work, we'll compress the sequences with seqdss
#   could also aggregate with dply later
seqtab( seqdss(mvad.seq), tlim = 0, format = "SPS" ) %>>%
  attr("freq") %>>%
  (
    data.frame(
      # appending "-end" is necessary for this to work
      sequence = paste0(
        gsub(
          x = names(.$Freq)
          , pattern = "(/[0-9]*)"
          , replacement = ""
          , perl = T
        )
        ,"-end"
      )
      ,freq = as.numeric(.$Freq)
      ,stringsAsFactors = FALSE
    )
  ) %>>%
  sunburst

Additional Reading

• Fast Company’s To Take on HBO and Netflix, Youtube Had to Rewire Itselffeaturing Kerry Rodden and this sequences sunburst.

• US Sirelines with a very nicely adapted Sequences Sunburst as a pedigree visualization for thoroughbred horses

• USERcycle blog post Behavior Flow

• Sequences Sunburst fork for Kaggle See Click Predict Fix

Thanks

Thanks so much

• Kerry Rodden for Sequences Sunburst
• Ramnath Vaidyanathan and RStudio for htmlwidgets
• all the contributors to R and JavaScript