Launch bettr app to explore and aggregate performance metrics

Usage

bettr(
  df = NULL,
  idCol = "Method",
  metrics = if (!is.null(df)) setdiff(colnames(df), idCol) else NULL,
  initialWeights = NULL,
  initialTransforms = list(),
  metricInfo = NULL,
  metricColors = NULL,
  idInfo = NULL,
  idColors = NULL,
  weightResolution = 0.05,
  bstheme = "darkly",
  appTitle = "bettr",
  bettrSE = NULL,
  addStopButton = TRUE,
  defaultWeight = 0.2,
  serverMode = FALSE,
  cacheVersion = NULL
)

Arguments

df

A data.frame in wide format. Should contain one column with the IDs of the entities to be compared, and one column for each metric to use for the comparison.

idCol

Character scalar, indicating the name of the column of df and/or idInfo that contains IDs of the entities to be compared (e.g., methods).

metrics

Character vector, indicating which of the columns of df that correspond to metrics of interest. Only metrics included here will be displayed.

initialWeights

Named numeric vector providing initial weights for each metric to use for aggregating them into a final score. Must contain one entry per metric included in metrics.

initialTransforms

Named list with initial values of transformation parameters for each metric. Each list entry should correspond to one metric, and take the form of a list with up to four elements, named:

* **flip**: Logical scalar; whether or not to flip the sign of the
    metric values. Defaults to `FALSE`.
* **offset**: Numeric scalar; offset to add to the (flipped)
    metric values. Defaults to `0`.
* **transform**: Character scalar; one of 'None', 'z-score',
    '\[0,1\]', '\[-1,1\]', 'Rank', 'Rank+\[0,1\]' or 'z-score+\[0,1\]',
    indicating which transform to apply to
    the metric values (after any flipping and/or adding the offset).
    Defaults to 'None'.
* **cuts**: Numeric vector or `NULL`; the cut points that will
    be used to bin the metric values (after the other transformations).
    Defaults to `NULL`.

Only values deviating from the defaults need to be explicitly specified,
the others will be initialized to their default values.

metricInfo

data.frame with annotations for metrics. Must have a column named 'Metric' identifying the respective metrics.

metricColors

Named list with colors used for columns of metricInfo. Should follow the format required for ComplexHeatmap heatmap annotations. The list can include an entry named 'Metric', which contains a named vector with colors to use for metrics.

idInfo

data.frame with annotations for entities. Must have a column named according to idCol identifying the respective entities.

idColors

Named list with colors used for columns of idInfo. Should follow the format required for ComplexHeatmap heatmap annotations. The list can include an entry named according to idCol, which contains a named vector with colors to use for entities.

weightResolution

Numeric scalar in (0,1), giving the resolution at which weights can be specified using the sliders in the interface.

bstheme

Character scalar giving the bootswatch theme for the app (see https://bootswatch.com/). Default 'darkly'.

appTitle

Character scalar giving the title that will be used for the app. Defaults to 'bettr'.

bettrSE

A SummarizedExperiment generated by assembleSE. If this is not NULL, df, metrics, initialWeights, initialTransforms, metricInfo, metricColors, idInfo and idColors arguments will be ignored and the information will be extracted from the SummarizedExperiment object.

addStopButton

Logical scalar. If TRUE (default), will add a button to stop the app (by calling shiny::stopApp).

defaultWeight

Numeric scalar between 0 and 1, giving the default weight to assign to each metric.

serverMode

Logical scalar. If TRUE, launches the app in server mode where users can upload JSON files (in bettr format). If FALSE (default), requires data to be provided via the df or bettrSE parameter.

cacheVersion

Character string or NULL (default). A version identifier for the browser cache. When specified, cached data from previous versions will be automatically invalidated. Useful for forcing cache refresh when deploying updates. Examples: "v1.0", "2024-01-15", or any string.

Value

A shiny application

Author

Charlotte Soneson, Daniel Incicau

Examples

df <- data.frame(Method = c("M1", "M2", "M3"), metric1 = c(1, 2, 3),
                 metric2 = c(3, 1, 2), metric3 = factor(c("a", "a", "b")))
initialTransforms <- list(metric1 = list(flip = TRUE, offset = 4))
metricInfo <- data.frame(Metric = c("metric1", "metric2", "metric3"),
                         Group = c("G1", "G2", "G2"))
idInfo <- data.frame(Method = c("M1", "M2", "M3"),
                     Type = c("T1", "T1", "T2"))
metricColors <- list(Group = c(G1 = "red", G2 = "blue"))
if (interactive()) {
    bettr(df = df, idCol = "Method",
    metrics = c("metric1", "metric2", "metric3"),
    initialTransforms = initialTransforms,
    metricInfo = metricInfo, metricColors = metricColors,
    idInfo = idInfo)
}