I have checked up the official webpage of Plotly, but how to plot the following task still escapes me:
id <- c(rep(1,5), rep(2,5), rep(3,5), rep(4,5))
t <- rep(seq(50,75,length.out=5), 4)
x <- runif(20) + sin(t)
y <- rnorm(20) + cos(t)
gender <- c(rep("F",10), rep("M",10))
smoke <- c(rep("Y",5), rep("N",10), rep("Y",5))
DATA <- data.frame(ID, t, x, y, gender, smoke)
fig <- plot_ly(DATA, x = ~t, y = ~y, z = ~x, .......)
Suppose I have 4 groups of patients (grouped by 2 factors, Female/Male and Smokers/Non-smokers), each associated with 5 observations $(x_i, y_i)$ along the timestamps $t_i$. So I need to draw a 3D line plot $${(t_i, x_i, y_i)}_{i=1}^{i=5}$ for each patient, but all in one plotting canvas. If I want to represent genders by red/blue, smokers by solid and non-smokers by dash lines, and specify these in the legend, what should I do (ideally using R)?
The kind of 3D plot you have in mind can be achievd like so:
library(plotly)
id <- c(rep(1,5), rep(2,5), rep(3,5), rep(4,5))
t <- rep(seq(50,75,length.out=5), 4)
x <- runif(20) + sin(t)
y <- rnorm(20) + cos(t)
gender <- c(rep("F",10), rep("M",10))
smoke <- c(rep("Y",5), rep("N",10), rep("Y",5))
DATA <- data.frame(id, t, x, y, gender, smoke)
col_gender <- c(M = "red", F = "blue")
lt_smoke <- c(Y = "solid", N = "dash")
sym_id <- c(`1` = "circle", `2` = "square", `3` = "diamond", `4` = "cross")
fig <- plot_ly(DATA,
x = ~x, y = ~y, z = ~t, symbol = ~id, color = ~gender, linetype = ~smoke, type = 'scatter3d', mode = 'lines+markers',
line = list(width = 6),
marker = list(size = 3.5, cmin = -20, cmax = 50),
colors = col_gender,
linetypes = lt_smoke,
symbols = sym_id)
fig
In case of more patients the best option is to map id
on color and additonally group by id
using transform groupby
library(plotly)
id <- c(rep(1,5), rep(2,5), rep(3,5), rep(4,5), rep(5,5), rep(6,5), rep(7,5), rep(8,5))
t <- rep(seq(50,75,length.out=5), 8)
x <- runif(40) + sin(t)
y <- rnorm(40) + cos(t)
gender <- c(rep("F",10), rep("M",10), rep("F",10), rep("M",10))
smoke <- c(rep("Y",5), rep("N",10), rep("Y",5), rep("Y",5), rep("N",10), rep("Y",5))
lt_smoke <- c(Y = "solid", N = "dash")
sym_id <- c(M = "circle", F = "square")
fig <- plot_ly(DATA,
x = ~x, y = ~y, z = ~t, symbol = ~gender, color = ~id, linetype = ~smoke, type = 'scatter3d', mode = 'lines+markers',
line = list(width = 6),
marker = list(size = 3.5, cmin = -20, cmax = 50),
linetypes = lt_smoke,
symbols = sym_id,
transforms = list(
list(
type = 'groupby',
groups = ~id)
))
fig