Stat200 lab3

Statistics 200: Lab 3 (Friday 30 Janurary 98)

Today's tasks:
Fancier graphics: beyond the defaults. Multiple plots per page, split screens, "graphics frames".

Splus has some very powerful built-in graphics tools. A great feature of Splus is that you can construct your own graphics functions and customize your plots to do exactly what you want. Today we shall construct our own graphical tool.

We have saved some Census data in the Stat200 library (in the section called nhcensus. You should attach that section of the library to get access to the data set nhages, which is a list. ( See HELP if you have forgotten about libraries.)

The data gives counts for New Haven population, broken down by race, sex, and agegroups. (Here is the html version of the file, as it came from the Census lookup page. Don't try to copy from the html; use the Stat200 library!)

This week I want you to write a series of functions that generate population pyramids (similar to this example for Hartford County) for the New Haven data.

Problem 1

Write a function to construct a population pyramid for white population of the town. The plot should represent percentages of the population for the various age groups, with females on one side of the central vertical axis and males on the other. Hint: The barplot command will be very useful (Look at the various optional arguments to barplot. You should find ways to tip a barplot on its side, to make bars stick out in two directions, and other exciting variations!)

Problem 2

Improve your function from Problem 1:

make sure the ages increase from the bottom of the pyramid.
make the axis show percentages of population increasing in both directions away from the central vertical axis.
add a title
fill in any gaps between the horizontal bars
make sure any labels fit on the page

Some hints: The par command is used to change many of the graphical parameters in Splus, it has many arguments. The Splus help file for par is overpowering and we have included a more simple form of the file here

(see remarks on the dreaded par()).

You will need the commands: axis and pretty to do some of Problem 2; Check here for more hints.

Problem 3

Improve your function from Problem 2: group the population into 5 year age ranges, 0-4 years, 5-9 years, and so on. Redraw the pyramid. Hint: ?paste, ?rep, ?tapply. Also, try this little experiment, to see one way of collapsing categories:
indices <- rep(1:6,c(3,4,2,6,2,2))
full <- 1:length(junk)
squished <- tapply(full,indices,sum)
Now use paste() to make suitable labels for the squish values.

Problem 4

Write a function to draw three little population pyramids in a row: white, black, hispanic. Have a title on each pyramid, and a main title for the plot. Hint: To divide the graphics screen into sections use the par command (the mfrow argument) or the split.screen command, I warn you the split.screen command can get confusing.