Original data source: U.S. Census Bureau,

<http://www.census.gov/popest/archives/pre-1980/PE-11.html>

Note: Subsequent to restructuring of the U.S. Census Bureau website, the original URL is no longer valid. However, the data files appear to be available at 

<https://www2.census.gov/programs-surveys/popest/tables/1900-1980/national/asrh/>


Steps: 

1. Download CSV files (e.g. using wget)

2. Preprocess using a shell script:

for f in PE-11-19??.csv ; do
    g=${f/PE-11-/}
    echo $g
    grep ^[0-9] $f | sed -e 's/,//g' | sed -e 's/"/ /g' | sed -e 's/ /,/g' | sed -e 's/,,,/,/g' | sed -e 's/,,//g' | cut -d, -f1,3,4 > $g
done

3. Convert to an R array:

readData <-
    function(file)
{
    ans <-
        read.table(file,
                   header = FALSE,
                   sep = ",",
                   row.names = 1,
                   col.names = c("Age", "Male", "Female"),
                   as.is = TRUE)
    ans <- as.matrix(ans)
    upto74 <- ans[1:75, ]
    above74 <- colSums(ans[-(1:75), , drop = FALSE])
    ans <- rbind(upto74, `75+` = above74)
    names(dimnames(ans)) <- c("Age", "Sex")
    ans
}


Year <- 1900:1979

foo <- lapply(paste(Year, "csv", sep = "."), readData)

USAge.table <-
    array(do.call(c, foo),
          dim = c(dim(foo[[1]]), length(Year)),
          dimnames = c(dimnames(foo[[1]]), list(Year = as.character(Year))))

save(USAge.table, file = "USAge.table.rda", compress = TRUE)

4. Make alternative data frame version (omit 75+ group)

USAge.df <- as.data.frame.table(USAge.table[1:75, , ])
##levels(USAge.df$Age)[nlevels(USAge.df$Age)] <- "75"
USAge.df$Age <- as.numeric(as.character(USAge.df$Age))
USAge.df$Year <- as.numeric(as.character(USAge.df$Year))
USAge.df$Population <- USAge.df$Freq / 1e6
USAge.df$Freq <- NULL

save(USAge.df, file = "USAge.df.rda", compress = TRUE)