Graphics Reference
In-Depth Information
world. his applet is available with Java Docs to other federal agencies and may be
generally available for other developers as well. In fact, as already stated, the USU
WNV Web-based micromap application described in Sect.
.
.
is based on the NCI
Java code. heplaces in the Java code to make changes for other applications are iso-
lated. he graphical user interface (GUI) prompts, options, and subsequent database
access need to be modified for boundary files if regions other than the
US states
and underlying counties are used. Similarly, the GUI prompts, options, and database
access need to be modified for different data. his is straightforward for experienced
Javaprogrammers.TwoGMUstudentswereabletomakethemodificationtoobtain
EPA'stoxic release inventory data over the Web,andmainly one USUstudent did the
adaptation for the WNV micromap server.
Implementations of LM plots are appearing in new settings. he Economic Re-
search Service (ERS) of the USDA now has the capability to produce LM plots using
sotware called Pop Charts. Micromaps recently were created for French régions and
départements andmaybeincludedintheexploratory spatial analysis packageGeoXP
(http://w
.univ-tlse
.fr/GREMAQ/Statistique/geoxppageR.htm).
Discussion
1.6
In this chapter, we have demonstrated the use of interactive linked micromap plots
forthedisplayofgeographicallyreferencedstatisticaldata.Someoftherestrictionsof
choropleth maps donot apply for LMplots. It is possible to displaymultiple variables
at a time in LM plots, provide summary statistics, and maintain the exact ranking of
different subregions. he recent use of interactive LM plots on federal (and other)
Web sites and their use in geographic publications are encouraging indicators for
their widespread acceptance. We want to finish with a comparison of LM plots with
ConditionedChoroplethMaps(CCmaps),introducedinCarretal.(
b)andCarr
etal.(
),andTrellisGraphics,basedonideasusedinCleveland(
)andfurther
discussed in Sect.
.
in Venables and Ripley (
).
Carr et al.(
b)developed conditioned choropleth mapsas a waytoshowthree
variables using choropleth maps.he basic idea was to usethe conditioning method-
ology described by Cleveland et al. (
) to partition the regions in a map in a
set of panels containing partial maps. Figure
.
provides a CCmaps example based
on the data from Fig.
.
. he top slider in Fig.
.
shows boundaries to convert the
soybean production (in bushels) per state into three color classes for use in
a choropleth map. he bottom slider in Fig.
.
shows acreage boundaries for par-
titioning states into the columns of the
set of map panels. he right slider in
Fig.
.
showstheyieldboundaries forpartitioning statesintorowsofthemappanels.
States with highacreageandhighyieldappearinthetoprightpaneland,asexpected,
have a high number of bushels. he let column of panels highlights states with low
acreage. Maryland and Pennsylvania are in the middle and high yield classes and are
in the middle class in terms of bushels. All the other states in the let column are in
