Graphics Reference
In-Depth Information
range of capabilities includes creating interactive tables and linked micromaps. Ex-
periences with nViZn, its advantages and current problems, and its capabilities for
the display of federal data are described in more detail in Jones and Symanzik ( ),
Symanzik and Jones ( ), Symanzik et al. ( ), and Hurst et al. ( ).
Whilethemainapplication ofmicromapsundernViZn wasintendedtobeaproof
of concept, based on the original EPA HAP data, the implementation of this applica-
tionisveryflexible.Whenadditionaldatafilesintheappropriateformatareprovided,
these data will be immediately usable within the nViZn application. he current ap-
plication uses data at the national (US), state, county, and census tract level.
he micromaps created under the nViZn application are dynamic. he user can
sort the regions in ascending or descending order with respect to the six statistical
criteria minimum, mean, maximum, first quartile, median, and third quartile of the
underlyingcensustractlevel.henumberofregionsthataredisplayedpermicromap
can be changed “on the fly” via a slider, so a user is not restricted to the perceptual
grouping of size five or less that was introduced in Sect. . and can experiment with
othergroupsizes.Micromapsandtables ofthedatacanbecreatedbyselectingaHAP
andUSstateinadrill-downmap.MultipleLMplotsortabulardisplayscanbeviewed
simultaneously.
henViZn micromap application only workstogether with nViZn. hemicromap
application can be obtained freely from the first author upon request. At this point,
SPSS no longer sells nViZn as a commercial product. However, the nViZn visualiza-
tion service is a free simple service-based Java servlet application, accessible via the
nViZn Web site mentioned above. Users pass GPL statements to the service via http
requests and get a graph in return.
Micromaps via Java and Other Statistical Packages
1.5.3
he NCI recently developed a Java application to make LM plots available to those
whoare not familiar with statistical packages. hemicromap producer must provide
region boundary files in the form of .gen (or .shp files in the near future) and data
files in the form of comma delimited (.csv) files. he interactive dialog enables the
selection of plot types and the selection of columns from the data file to associate
with the specific plot. For example, a dotplot with precalculated confidence bounds
requiresthree variables. Atthetimeofthiswriting, thesotwareisstill beingtestedin
the classroom at George Mason University (GMU) and at other NCI-approved sites
priortogeneral release.hesotware will mostlikely beavailable tothe publicbefore
this chapter is published.
Mostmajorstatistical packagesareflexible enoughtoproduceLMplots.hebasic
issuesareavailability ofboundary files,convenience ofproduction,quality ofappear-
ance, and output format. For example, LM plots appeared some time ago in Wilkin-
son( ),introducing thegrammarof graphics. SPSSversion . (http://spss.com/
spss/) provides an implementation of this grammar, so this should make the produc-
tion of LM plots easy for those who are familiar with the grammar of graphics.
he NCI LM plot applet in the State Cancer Profiles Web site, discussed in more
detail in Sect. . . , provides LM plots to health planners in countries around the
Search WWH ::




Custom Search