StreamGraph

An R Version of the D3.js Streamgraph Example

Leveraging the R package streamgraph which requires htmlwidgets, you can build the same streamgraph in R in about 3 lines of R code.

The package has some limitations, however. For example, no chart title seems to be possible on the interactive version - at least as an explicit parameter.

However, it is less complex to implement than the d3 version because the low level steps are abstracted by the single function call in R.

library(tidyverse)
library(streamgraph)
library(kableExtra)

After reading the raw data file into csv format, pivot it to a required long skinny format. Then because the index needs to recognized as a year value, add 1900 to the index to make it interpretable as a year.

yy = read_csv("ue_industry.csv")

yy %>% pivot_longer(!index, names_to = 'industry', values_to = "count") -> ts2

#
# In order to convert the X-axis to a year parameter 
# as required by the R streamgraph package.  We shift the index by a constant.
# ------------------------------------------------------------------------------
ts2$index = ts2$index + 1900

The resulting streamgraph below replicates the d3 result reasonably well.

pp2 = streamgraph(ts2, key= "industry",
                   value="count",
                  date="index", height="500px", width="800px", order = "inside-out")

pp2

Lastly, representative pivot table data is shown below.

head(ts2, n=20) %>% kable() %>% kable_styling(bootstrap_options = c("hover", "striped"))

index	industry	count
1900	Agriculture	154
1900	Business services	655
1900	Construction	745
1900	Education and Health	353
1900	Finance	228
1900	Government	430
1900	Information	125
1900	Leisure and hospitality	782
1900	Manufacturing	734
1900	Mining and Extraction	19
1900	Other	274
1900	Self-employed	239
1900	Transportation and Utilities	236
1900	Wholesale and Retail Trade	1000
1901	Agriculture	173
1901	Business services	587
1901	Construction	812
1901	Education and Health	349
1901	Finance	240
1901	Government	409