ez.scatterplot: scatter plot with lattice

In jerryzhujian9/zmisc: ez Zhu's miscellaneous functions

scatter plot with lattice

Description

scatter plot with lattice

Usage

ez.scatterplot(
  df,
  cmd,
  lmline = TRUE,
  loess = FALSE,
  model = c("lm", "lmrob", "lmRob", "rlm"),
  type = c("spartial", "partial"),
  scale = FALSE,
  rp = TRUE,
  rp.x = 0.025,
  rp.y = 0.05,
  se = TRUE,
  layout = NULL,
  rp.size = 14,
  line.width = 3,
  point.size = 14,
  x.axis.size = 16,
  y.axis.size = 16,
  x.lab.size = 18,
  y.lab.size = 18,
  title.size = 20,
  legend.size = c(0, 14),
  line.color = "#BE1B22",
  line.style = 1,
  loess.color = "dark grey",
  loess.width = 3,
  loess.style = 2,
  point.color = "#0086B8",
  point.shape = 16,
  point.alpha = 0.9,
  colors = ez.palette("Zhu"),
  shapes = c(16, 17, 15, 3, 7, 8),
  ylab = NULL,
  xlab = NULL,
  xbreaks = NULL,
  ybreaks = NULL,
  xlimits = NULL,
  ylimits = NULL,
  title = NULL,
  zlab = NULL,
  legend.box = FALSE,
  legend.position = "top",
  legend.direction = "horizontal",
  hack = FALSE,
  print2scr = TRUE,
  ...
)

Arguments

df	data frame
cmd	like "y~x+a+b", "y~x+a+b|z", "y~x+a+b||z", "y~x+a+b|||z"where y x are continous, z discrete (| one regression line, || multiple regression lines by levels of z–gives interaction p value), +a+b optional for covariates residualization y~x+a+b|||z plots separated by z groups, y~x+a+b||||z plots with and without z grouping
lmline	T/R lmline (for all data points), particularlly useful when separate lines are plotted for each group
loess	T/R adds a loess fit (uses panel.loess, the same as type = c("smooth"))
model	one of c('lm', 'lmrob', 'lmRob', 'rlm'), robustbase::lmrob–MM-type Estimators; robust::lmRob–automatically chooses an appropriate algorithm. one or more, 'lm' will always be included internally, even if not specified
type	one of c('partial','spartial'), or c('p','sp'), plot partial or semi-partial correlation. Only applicable when covar exists
scale	when having covariates, z transform residuals
rp	show r (signed) and p values
rp.x	r p values x position (0-1, relative to top left, big–>right), ignored if rp=FALSE.
rp.y	r p values y position (0-1, relative to top left, big–>up), ignored if rp=FALSE.
se	standard error of linear regression line
rp.size	r p values font size, ignored if rp=FALSE
point.size	if less point, one can increase size
legend.size	c(0,12) the first number 0 controls the legend title, 0=hide; the second number controls legend.key.size, legend.text
line.color	only applicable when not auto varied, regression line color.
point.color	only applicable when not auto varied
point.shape	only applicable when not auto varied
point.alpha	if overplot for points, one can reduce alpha
ylab	y label. NULL
xlab	x label. NULL
xbreaks	seq(from=95, by=0.25, to=97). xbreaks does not require xlimits set
xlimits	c(0,1). xlimits does not require xbreaks set
zlab	z legend label/title, only applicable when there is z provided. NULL
legend.box	box of legend, T or F
legend.position	legend position 'none' (I think this is not natively supported, but works) "left", "right", "top", and "bottom" c(x,y,two-element fractional numeric vector) c(0,0) corresponds to the "bottom left" and c(1,1) corresponds to the "top right" position, but within the plot box.
legend.direction	horizontal or vertical
hack	T/F if more than 1 model specified, plot all in one plot
...	other parameters passed to xyplot. E.g., jitter.x=T, jitter.y=T, factor=1, amount=NULL see jitter layout=c(col,row)–a bit weird. lattice orders panels from the bottom up. To get what you want either use as.table = TRUE or reverse the ordering in index.cond. eg, as.table=F, index.cond=list(c(3,2,1))

Value

a lattice plot, can use update() to change the plot

Note

specify auto through param colors, shapes

jerryzhujian9/zmisc documentation built on March 9, 2024, 12:49 a.m.