ggplot aestethics in ts_plot and issue saving plot

[mpaulasg]

Hi Mikko,

Thank you for this package! I'm able to obtain the T-S plot successfully, but I'm now trying to change the size and colour of axis titles, etc, and I'm not able to do this. I've tried adding this change inside ts_plot and outside, but no success. Is this possible?

Another issue is that I'm trying to save the plot using ggsave, but instead I obtained an empty plot. I can save other plots using the exact same code, so I guess it should be an issue with this function. See an example below.

Thank you in advance!
Paula

library(PlotSvalbard)
library(here)

#Setting some graphical details

color_cross <- c("Inshore" = "#6C809A", "Midshelf" = "#FAAA8D", "Offshore" = "#98473E")

#Create data

CTD <- data.frame(Habitat = c("Inshore", "Inshore", "Inshore", "Midshelf", "Midshelf","Midshelf", "Offshore",
"Offshore","Offshore"),
Depth = c(1, 2, 3,1, 2, 3,1, 2, 3 ),
Temp = c (23, 22.5, 21, 22, 22, 24.5, 24, 25, 24.5),
Sal = c(35.5, 35.8, 35, 34.5, 34.6, 34.6, 34, 34.2, 34))

#Plot TS

ts <- ts_plot( CTD,
temp_col = "Temp",
sal_col = "Sal",
WM = NULL,
color_wmpoly = NULL,
xlim = c(33,36),
ylim = c(20,26),
color = "Habitat",
zoom = TRUE,
nlevels = 1,
color_isopyc = "grey90",
symbol_shape = 16,
symbol_size = 3,
symbol_alpha = 0.6,
color_scale = color_cross,
color_var_name = NULL,
margin_distr = TRUE, #this one shows Kernel density
margin_width = 0.15,
margin_height = 0.2,
plot_data = TRUE,
base_size = 10

)

ggsave(ts, file=here::here("graphs", "ts_cross.png"),
height = 22, width = 25, unit = "cm" ) #Doesn't work

[MikkoVihtakari]

Hi Paula,

Thanks for the bug-report. Try rather ggOceanPlots. It is more updated (and development of PlotSvalbard has ended). It seems that the marginal plot code in the ts_plot() function is outdated. It plots to grid instead of returning a grid (or ggplot2) object. I'll update the code when I'll have time (I have solved the issue in more recent scripts which I have not shared here). In the meanwhile you can either not use marginal distributions (set margin_distr to FALSE) or by-pass the problem by printing inside the png() device. The latter option does not let you control the axes labels using ggplot2 functions (because ts_plot() does not return a ggplot2 object when margin_distr = TRUE; this is a clear design issue and a bug that can be fixed).

library(ggOceanPlots)
library(ggplot2)

#Setting some graphical details

color_cross <- c("Inshore" = "#6C809A", "Midshelf" = "#FAAA8D", "Offshore" = "#98473E")

#Create data

CTD <- data.frame(Habitat = c("Inshore", "Inshore", "Inshore", "Midshelf", "Midshelf","Midshelf", "Offshore",
                              "Offshore","Offshore"),
                  Depth = c(1, 2, 3,1, 2, 3,1, 2, 3 ),
                  Temp = c (23, 22.5, 21, 22, 22, 24.5, 24, 25, 24.5),
                  Sal = c(35.5, 35.8, 35, 34.5, 34.6, 34.6, 34, 34.2, 34))

png("test.png", height = 22, width = 25, unit = "cm", res = 300)
ts_plot(CTD,
        temp_col = "Temp",
        sal_col = "Sal",
        WM = NULL,
        color_wmpoly = NULL,
        xlim = c(33,36),
        ylim = c(20,26),
        color = "Habitat",
        zoom = TRUE,
        nlevels = 1,
        color_isopyc = "grey90",
        symbol_shape = 16,
        symbol_size = 3,
        symbol_alpha = 0.6,
        color_scale = color_cross,
        color_var_name = NULL,
        margin_distr = TRUE, #this one shows Kernel density
        margin_width = 0.15,
        margin_height = 0.2,
        plot_data = TRUE,
        base_size = 10
) 
dev.off()

ts <- ts_plot(CTD,
        temp_col = "Temp",
        sal_col = "Sal",
        WM = NULL,
        color_wmpoly = NULL,
        xlim = c(33,36),
        ylim = c(20,26),
        color = "Habitat",
        zoom = TRUE,
        nlevels = 1,
        color_isopyc = "grey90",
        symbol_shape = 16,
        symbol_size = 3,
        symbol_alpha = 0.6,
        color_scale = color_cross,
        color_var_name = NULL,
        margin_distr = FALSE, #this one shows Kernel density
        margin_width = 0.15,
        margin_height = 0.2,
        plot_data = TRUE,
        base_size = 10) +
  ggplot2::theme(axis.title = element_text(size = 20, color = "blue"))

ggsave(ts, file = "test2.png",
       height = 22, width = 25, unit = "cm" ) # Works

[mpaulasg]

Hi Mikko,

So sorry for my late reply!

Thank you, I'm now using ggOceanPlots. I want to keep the Kernel density graphs so unfortunately, I can't set the marginal distributions off. However, I just changed the letter sizes with base_size and that's good enough for my figures.

Thanks again!

Paula

[MikkoVihtakari]

Thanks. Let's just keep this open as a reminder that I'll need to look into this one day when I have time.