Under Construction
likersection is an R package that mines bivariate insights from survey data.
It does this by aggregating a Likert-valued quantity (such as a satisfaction
measure) with respect to pairs of explanatory variables (such as demographic
measures) in a two-dimensional bubble plot. Likert-valued data are tricky to
analyze because they are partly ordinal ("Strongly Agree" to "Strongly
Disagree" in a finite number of notches) and partly not (answering "Don't Know"
or leaving the question blank). By defining a likelihood function aware of this
complexity, likersection is able to sort the N-choose-2 possible aggregations
by an Akaike Information Critierion, answering the question "Which intersection
of two explanatory variables best expresses the contrast in this quantity we're
interested in?"
likersection then visualizes such insights as a two-way bubble plot. The size
of a bubble denotes the number of observations for the corresponding subset of
the bivariate sample space; each bubble is made of a colored wedge (denoting
the mean value of the continuous quantity) and a gray wedge (denoting the
proportion of Don't Know/null/blank responses).