0

I have been trying to run the script of this Wikipedia chart showing US unemployment.

The data is from http://download.bls.gov/pub/time.series/ln/ln.data.1.AllData & http://download.bls.gov/pub/time.series/ln/ln.series

    cat("Loading table -- might take some time\n");
    u <- read.table("ln.data.1.AllData", header=T, fill=T)
    u$time <- u$year + (as.numeric(u$period) - 1) / 12

    cat("Processing -- might take some time\n");
    u1 = subset(u, series_id == "LNS13025670")
    u2 = subset(u, series_id == "LNS14023621")
    u3 = subset(u, series_id == "LNS14000000")
    u4 = subset(u, series_id == "LNS13327707")
    u5 = subset(u, series_id == "LNS13327708")
    u6 = subset(u, series_id == "LNS13327709")

par(family="Times")
par(bty = "n")
plot(
    0,
    main = "Measurement of unemployment",
    ylim = c(0,18),
    xlim = c(1950, 2010),
    xlab = "Year",
    ylab = "Percentage",
    las = 1
);

grid()

pal = rainbow(8)
lines(value ~ time, u6, col=pal[6])
lines(value ~ time, u5, col=pal[5])
lines(value ~ time, u4, col=pal[4])
lines(value ~ time, u3, col=pal[3])
lines(value ~ time, u2, col=pal[2])
lines(value ~ time, u1, col=pal[1])

legend(
    "topleft",
    rev(c(
        "U1: Percent Of Civilian Labor Force Unemployed 15 Weeks and over",
        "U2: Unemployment Rate - Job Losers",
        "U3: Unemployment Rate",
        "U4: All of U3, plus discouraged workers",
        "U5: All of U4, plus marginally attached workers",
        "U6: All of U5, plus total employed part time for economic reasons"
    )),
    col = rev(pal[1:6]),
    bty = 'n',
    lty = 1
)

dev.copy(svg, "US Unemployment measures.svg", width=8, height=6)
dev.off()

Despite it's the unmodified source code form Wikimedia Commons, the lines are all bogus:

PNG

What's wrong with the R script?

Is it because u1-u6 are falsely being interpreted as factors?

Luis Paganini
  • 171
  • 9
  • If you're unsure about the structure of your data, call `str` on it to see if those are indeed factors or characters – camille Jun 19 '18 at 15:27
  • 1
    Those first two lines are some of the most misleading lines I've seen in a script in a while. – Dason Jun 19 '18 at 15:38
  • In case you didn't understand @Dason comment, see https://stackoverflow.com/a/6150337/2761575 – dww Jun 19 '18 at 16:21
  • @Dason The code I've copied started only with "if (F) {", so I've had to invent something. But it's indeed very funny. – Luis Paganini Jun 19 '18 at 19:04
  • @LuisPaganini To be blunt - nobody here cares about the code you copied. We care about the code you put here. If you want that code block to be run for your example then you don't need to have those two lines at all and should be removed. – Dason Jun 19 '18 at 19:05

1 Answers1

1

Just glancing at the raw data, there's an issue with your code:

u$time <- u$year + (as.numeric(u$period) - 1) / 12

But the period column has values like 'M01', 'M02', 'Q01', Q02'. Since that column contains characters, read.table converts it to factors by default (you can turn that off). Calling as.numeric on something like 'M01' will just return the numerical, ordinal value of the factor.

David Klotz
  • 2,401
  • 1
  • 7
  • 16