The statistics behind the bio.viz biomarker charts. A chart never computes a test: it hands over a table with one row per participant and receives the estimates, intervals, p-value, method name and counts that R worked out.
Design
Each statistic is a thin wrapper that fixes the inputs and the shape of the answer around a function from the stats or survival package. Nothing is reimplemented, and the statistics use nothing from any other package, so the same source runs in a desktop session, in R in the browser and on a server.
The tests the design names are imported here, once, for the functions that
wrap them: stats::t.test(), stats::wilcox.test(), stats::aov(),
stats::kruskal.test(), stats::cor.test(), stats::chisq.test(),
stats::fisher.test(), stats::p.adjust(), stats::lm() and
stats::loess() from stats;
survival::Surv(), survival::survdiff(), survival::survfit() and
survival::coxph() from survival.
Widgets
A widget draws a bio.viz chart from R. It computes the chart's statistics
with the functions above when it is made and stores them in the page, so a
saved page shows them with no R and no network: see
Widget_GroupComparison(), Widget_AssociationScatter(),
Widget_CorrelationMatrix(), Widget_BiomarkerScreen(), Widget_CrossTab()
and Widget_StratifiedSurvival(). The widgets are the only part of the
package that uses htmlwidgets.
Figures
Each chart has a static ggplot2 figure too, Visualize_<Chart>(), from the
same settings and with the same statistics: see
Visualize_GroupComparison() and the functions beside it. ggplot2 is
suggested, not imported.
Status
Version 0.2.0 is being prepared: the statistics functions, a synthetic biomarker study with known planted effects, a widget for each of bio.viz's six charts, and from R a static figure, a statistics table, RTF and batch runs of chart specifications. Version 0.1.0 is the latest release.
Author
Maintainer: obot obot@users.noreply.github.com