Graphics Reference
In-Depth Information
Figure
.
.
Choropleth maps of the
Census of Agriculture, showing the variables soybean yield (in
bushels per acre) and acreage (in millions of acres) by state. he data represent the
US states where
soybeans were planted
all, there appears to be a positive correlation between these two variables since high
yields/high acreages and low yields/low acreages seem to appear in the same geo-
graphic regions. he correlation coe
cient between yield and acreage is only
.
,
suggesting departures from linearity that would be better revealed using scatterplots
or LM plots.
heLMplotsparadigmsortsthelistofregions basedontheirvalues ornamesand
partitions thelistintoperceptualgroupsofsizefiveorless,whichisfurtherdiscussed
in Sect.
.
. he micromap design assigns a distinct color to each region in a group.
he same color is used for plotting the region polygon in the micromap, the dot by
theregion name,andthe region values in oneormultiple statistical panels.Figure
.
provides a LM plot example with
US states and shows three statistical panels with
dotplots for three variables: yield, acreage, and production. his example is accessi-
ble at http://www.nass.usda.gov/research/gmsoyyap.htm. In fact, Fig.
.
shows the
LM plots of the same two variables as Fig.
.
, plus a third statistical panel for the
