# Graphics problems
#------------------

# Load the classroom data you saved in the Reading data exercises,
#  or just load() it from our website.

# Plot 1
#-------
# Create a histogram of the mathkind (mathematics score in
#  kindergarten) variable.  Add an xlab= to better explain
#  what the graph is about.


# Plot 2
#-------
# Create an empirical density plot of mathkind.

# Plot 3
#-------
# Create an empirical density plot of mathkind by sex using
#  the optional argument groups = sex.  It is probably a good
#  idea to suppress the points in the "rug", with the plot.points=FALSE
#  option. Remember to use auto.key so you can tell which curve is which.

# Plot(s) 4
#-------
# In xyplot(), the default it to plot points.  You can plot an
#  empirical "smoother" instead, with option type="smooth".  You
#  can overlay to "types" that use the same data with
#  type=c("p", "smooth").

#  Create a scatter-plot of the mathgain versus the kindergarten score.
#  Add a smoother curve and a ylab (y axis label).  

# Repeat this plot using type "r"
#  instead of "smooth" to add a reference (or "regression")
#  line.

# Plot 5
#-------
# Find the correlation between mathkind and mathgain.  The negative 
#  result is not surprising because mathgain is the grade 1 score minus
#  the kindergarten score.  Create a new variable math1 which is
#  the sum of mathkind and mathgain.

#  Plot math1 versus mathkind.  Try using the
#  optional argument aspect="iso" to ensure that a
#  unit change on the x axis corresponds to a unit change on the y axis.

# Plot 6
#-------
# Create a multi-panel scatterplot of math1 versus
#  mathkind with separate panels for males and females.

# Plot 7
#-------
# Create a multi-panel scatterplot of math1 versus
#   mathkind classified according to sex and minority.

# Plot 8
#-------
# Consider only the students in school 11.  xyplot has a subset option
#  that allows you to specify, for instance, subset=(schoolid==11).
#  Show that the study includes students from nine different classrooms
#  in that school.

#  Create a dot-plot of the mathgain by classroom for students
#  in school 11 only.  Label the x and y axis.


# Plot 9
#-------
#  Repeat the plot, joining the classroom averages by overlaying
#  type "a".  

# Plot 10
#--------
#  Finally, you can reorder the classrooms with the
#  reorder() function (which relevels a factor).
#  reorder(classid, mathgain) will
#  do the trick (the default function applied to mathgain is
#  the mean() ).  

#  Redo the plot reordering the classrooms according
#  to increasing mean gain and joining the classroom averages.

# Bonus plots
#------------
# You can reorder() the levels of a factor by other functions.  How
#   would the previous example change if we reordered by max(mathgain)?
#   by mean(ses)?