ggplot melted data sharing common aesthetics

I need to plot a graph from a melted data.table that mixes three geoms: geom_line, geom_col and geom_point. geom_line and geom_point share a the color aesthetics:

set.seed(1)  time <- 1:10  mag  <- 10:20  q_mag <- c("up", "up", "down")  l_a  <- sample(1:12, 10)  l_b  <- sample(8:20, 10)  p_a  <- sample(1:12, 10)  p_b  <- sample(8:20, 10)    dt <- data.table(time, mag, q_mag, l_a, l_b, p_a, p_b)  dt <- melt(dt, measure.vars = c("mag", "l_a", "l_b", "p_a", "p_b"))      ggplot(data = dt, aes (x = time)) +    geom_col( data = dt[variable %in% "mag"]            , aes( y    = value                 , fill = q_mag)) +        scale_fill_grey(start = .4) +        geom_line( data = dt[variable %in% c("l_a", "l_b")]              ,aes( y = value                  , color = variable)              ,size =1) +        geom_point( data = dt[variable %in% c("p_a", "p_b")]               ,aes( y = value                    ,color = variable                    ,size = variable                    ,shape = variable)) +      scale_shape_manual( name = element_blank()                       , labels = c("Point A", "Point B")                      , values = c("circle", "cross")    )+            scale_size_manual( name = element_blank()                      , labels = c("Point A", "Point B")                     , values = c(4,2)    )+            scale_color_manual( name = element_blank()                      , labels = c("Line A", "Line B", "Point A", "Point B")                      , values = c("red", "green", "red", "blue")    )  

This code plots the graph as expected:

However, I would like to have the legend presented in another way:

• Point A as a large red circle,
• Point B as a small blue cross,
• Line A as a straight red line (no point), and
• Line B as a straight green line (no point).

How do I accomplish the desired legend?

https://stackoverflow.com/questions/65930458/ggplot-melted-data-sharing-common-aesthetics January 28, 2021 at 11:04AM