# plotly Multiple Generates multiple plotly graphics, driven by specs in a data frame ```r plotlyM( data, x = ~x, y = ~y, xhi = ~xhi, yhi = ~yhi, htext = NULL, multplot = NULL, strata = NULL, fitter = NULL, color = NULL, size = NULL, showpts = !length(fitter), rotate = FALSE, xlab = NULL, ylab = NULL, ylabpos = c("top", "y"), xlim = NULL, ylim = NULL, shareX = TRUE, shareY = FALSE, height = NULL, width = NULL, nrows = NULL, ncols = NULL, colors = NULL, alphaSegments = 1, alphaCline = 0.3, digits = 4, zeroline = TRUE ) ``` ## Arguments - `data`: input data frame - `x`: formula specifying the x-axis variable - `y`: formula for y-axis variable - `xhi`: formula for upper x variable limits (`x` taken to be lower value) - `yhi`: formula for upper y variable limit (`y` taken to be lower value) - `htext`: formula for hovertext variable - `multplot`: formula specifying a variable in `data` that when stratified on produces a separate plot - `strata`: formula specifying an optional stratification variable - `fitter`: a fitting such as `loess` that comes with a `predict` method. Alternatively specify `fitter='ecdf'` to use an internal function for computing and displaying ECDFs, which moves the analysis variable from the y-axis to the x-axis - `color`: `plotly` formula specifying a color variable or e.g. `~ I('black')`. To keep colors constant over multiple plots you will need to specify an AsIs color when you don't have a variable representing color groups. - `size`: `plotly` formula specifying a symbol size variable or AsIs - `showpts`: if `fitter` is given, set to `TRUE` to show raw data points in addition to smooth fits - `rotate`: set to `TRUE` to reverse the roles of `x` and `y`, for example to get horizontal dot charts with error bars - `xlab`: x-axis label. May contain html. - `ylab`: a named vector of y-axis labels, possibly containing html (see example below). The names of the vector must correspond to levels of the `multplot` variable. `ylab` can be unnamed if `multplot` is not used. - `ylabpos`: position of y-axis labels. Default is on top left of plot. Specify `ylabpos='y'` for usual y-axis placement. - `xlim`: 2-vector of x-axis limits, optional - `ylim`: 2-vector of y-axis limits, optional - `shareX`: specifies whether x-axes should be shared when they align vertically over multiple plots - `shareY`: specifies whether y-axes should be shared when they align horizontally over multiple plots - `height`: height of the combined image in pixels - `width`: width of the combined image in pixels - `nrows`: the number of rows to produce using `subplot` - `ncols`: the number of columns to produce using `subplot` (specify at most one of `nrows,ncols`) - `colors`: the color palette. Leave unspecified to use the default `plotly` palette - `alphaSegments`: alpha transparency for line segments (when `xhi` or `yhi` is not `NA`) - `alphaCline`: alpha transparency for lines used to connect points - `digits`: number of significant digits to use in constructing hovertext - `zeroline`: set to `FALSE` to suppress vertical line at x=0 ## Returns `plotly` object produced by `subplot` ## Details Generates multiple `plotly` traces and combines them with `plotly::subplot`. The traces are controlled by specifications in data frame `data` plus various arguments. `data` must contain these variables: `x`, `y`, and `tracename` (if `color` is not an "AsIs" color such as `~ I('black')`), and can contain these optional variables: `xhi`, `yhi` (rows containing `NA` for both `xhi` and `yhi` represent points, and those with non-`NA` `xhi` or `yhi` represent segments, `connect` (set to `TRUE` for rows for points, to connect the symbols), `legendgroup` (see `plotly` documentation), and `htext` (hovertext). If the `color` argument is given and it is not an "AsIs" color, the variable named in the `color` formula must also be in `data`. Likewise for `size`. If the `multplot` is given, the variable given in the formula must be in `data`. If `strata` is present, another level of separate plots is generated by levels of `strata`, within levels of `multplot`. If `fitter` is specified, x,y coordinates for an individual plot are run through `fitter`, and a line plot is made instead of showing data points. Alternatively you can specify `fitter='ecdf'` to compute and plot emirical cumulative distribution functions. ## Examples ```r ## Not run: set.seed(1) pts <- expand.grid(v=c('y1', 'y2', 'y3'), x=1:4, g=c('a', 'b'), yhi=NA, tracename='mean', legendgroup='mean', connect=TRUE, size=4) pts$y <- round(runif(nrow(pts)), 2) segs <- expand.grid(v=c('y1', 'y2', 'y3'), x=1:4, g=c('a', 'b'), tracename='limits', legendgroup='limits', connect=NA, size=6) segs$y <- runif(nrow(pts)) segs$yhi <- segs$y + runif(nrow(pts), .05, .15) z <- rbind(pts, segs) xlab <- labelPlotmath('X₁₂', 'm/sec²', html=TRUE) ylab <- c(y1=labelPlotmath('Y1', 'cm', html=TRUE), y2='Y2', y3=labelPlotmath('Y3', 'mm', html=TRUE)) W=plotlyM(z, multplot=~v, color=~g, xlab=xlab, ylab=ylab, ncols=2, colors=c('black', 'blue')) W2=plotlyM(z, multplot=~v, color=~I('black'), xlab=xlab, ylab=ylab, colors=c('black', 'blue')) ## End(Not run) ``` ## Author(s) Frank Harrell

plotlyM function