Part 2 - mpg scatterplot - applied to average colour instagram by hour jan 1-31, 2016

Pontifications

Part 1
I realized that adding the colourname on 100K rows is an expensive i.e. slow operation. Therefore instead of creating a visualization with the colournames in part 2, in part 2 I am really just going to create a CSV file with the colournames added as an additional column that I can reuse for many visualizations without having to repeat the expensive colourname lookup operation.
Here’s the relevant code from https://github.com/rtanglao/2016-r-rtgram/blob/master/part2-create-csv-ig-van-average-colour-jan2016.R:

tc <- function(x) {
  return (head(color.id(x), n = 1))
}
 csv_url = "https://raw.githubusercontent.com/rtanglao/2016-r-rtgram/master/JANUARY2016/january2016-ig-van-avgcolour-id-mf-month-day-daynum-unixtime-hour.csv"
    average_colour_ig_van_jan2016 = read_csv(csv_url)
    print(average_colour_ig_van_jan2016)
    # the following line of code doesn't work because mutate just does the first value of the column
    # average_colour_ig_van_jan2016_with_colourname = mutate(average_colour_ig_van_jan2016, colourname = tc(colour))
   average_colour_ig_van_jan2016_with_colourname =  average_colour_ig_van_jan2016 %>% 
       rowwise() %>% 
       mutate(colourname = tc(colour))
    write_csv(average_colour_ig_van_jan2016_with_colourname,
              "january2016-ig-van-avgcolour-id-mf-month-day-daynum-unixtime-hour-colourname.csv")

%>% is like the unix pipe operator
rowwise() is required to repeat the function call on each row, without rowwise() you’ll just repeat it row 1 repeatedly i.e. get the same value each time, the value of tc applied to the first row !
mutate adds the colourname column by computing the colourname from the hexadecimal colour using the color.id function from the plotrix package
This RScript was invoked as follows:

cd JANUARY2016
Rscript ../part2-create-csv-ig-van-average-colour-jan2016.R

In Part 3, we’ll create a visualization using the colour names.