gainsChartBarGraph: Produce a quantile bar graph based on gains chart(s)

In dnegrey/miscTools: R package with miscellaneous tools

Description

gainsChartBarGraph produces a bar graph based on a mt_gainsChart object. When two such objects are supplied, the function produces a combined bar graph, where the first object (xb) is labeled as "Build" and the second object (xv) is labeled as "Validate".

Usage

gainsChartBarGraph(xb, xv = NULL, cb = "#009DDC", cv = "#77BF30",
  xLabel = "Model Quantile", yLabel = "Mean Outcome",
  yType = ifelse("binary" %in% class(xb), "pct", "dlr"),
  yDigits = ifelse("binary" %in% class(xb), 1, 0))

Arguments

xb	mt_gainsChart object
xv	mt_gainsChart object
cb	character string; fill color for xb bars (valid color)
cv	character string; fill color for xv bars (valid color)
xLabel	character string; x-axis label
yLabel	character string; y-axis label
yType	character string; y-axis format type; valid values are "int", "dlr" and "pct"
yDigits	non-negative integer value indicating the number of decimal places to show when hovering over the bars

Value

A plot_ly visualization of a mt_gainsChart object, showing the average Y value (avgY for "binary"/"continuous"; perFreqYC for "combined") by quantile ( yhatBin).

See Also

gainsChart, plot_ly

Examples

# pull in sample scored data frame
x <- modelSampleScored
# filter to validation subset
x <- split(x, x$ValidateFlag)
# gains chart binary
y <- lapply(x, function(z){gainsChart(z$TargetFlag, z$pTargetFlag)})
# bar graph
gainsChartBarGraph(y[[1]], y[[2]])
# gains chart continuous
y <- lapply(x, function(z){
    gainsChart(z[z$TargetFlag, "TargetValue"], z[z$TargetFlag, "pTargetValue"])
})
# bar graph
gainsChartBarGraph(y[[1]], y[[2]])
# gains chart combined
y <- lapply(x, function(z){
    gainsChart(z$TargetFlag, z$pTargetFlag*z$pTargetValue, z$TargetValue)
})
# bar graph
gainsChartBarGraph(y[[1]], y[[2]])

dnegrey/miscTools documentation built on May 3, 2019, 2:57 p.m.