Formatting Density of States Plots via ggplot2: Not Sure where to start

[hguthrie]

Hello,
I am trying to generate Density of States plots for some materials that I have data for but I am unsure how to use ggplot2 for this. Here is a picture of how the data is broken up. The x-axis is basically the Energy in eV and each of the other columns corresponds to the grouping of the data. I am trying to get each density plot to overlay onto each of the others but I can't figure out how to do that. All of the examples I have found are for data that also has a category column associated with it but in this case my different categories are the different y-value sets.
Any guidance is appreciated!

[FJCC]

I think you need to pivot your data to a longer format using the pivot_longer() function from the tidyverse. If your data frame is named DF, try

DF_lng <- DF |> pivot_longer(cols = TDOS:Cl..s., names_to = "state", values_to="DOS")

There may well be a mistake in that, since I don't have data to test it with.
With the longer format, you can set the fill or color aesthetic of ggplot to follow the "state" column.

[hguthrie]

If I were to pivot the data then how do I specify which values are used for which axis in ggplot, because wouldn't the columns then end up without names? And each row is named?

[joels]

Is this what you had in mind:

library(tidyverse)
library(scales)
theme_set(theme_classic(base_size=14))

# Fake data
set.seed(3)
DF = tibble(E..eV. = seq(-30, -10, length.out=10),
            TDOS = runif(10, 0, 0.5),
            Ru..s. = runif(10, 0, 0.5),
            Cl..s. = 1 - TDOS - Ru..s.)

DF.long = DF %>% 
  pivot_longer(cols = TDOS:Cl..s., names_to = "state", values_to="DOS") %>%
  group_by(E..eV.) %>% 
  mutate(ypos = cumsum(DOS) - 0.5*DOS) %>% 
  ungroup() 

yl = DF.long %>% 
  filter(E..eV.==min(E..eV.)) %>% 
  select(-c(E..eV., DOS)) %>% deframe()
yr = DF.long %>% 
  filter(E..eV.==max(E..eV.)) %>% 
  select(-c(E..eV., DOS)) %>% deframe()

DF.long %>% 
  ggplot(aes(x=E..eV., ypos, height=DOS, fill=state)) + 
    geom_tile(colour="white") +
    geom_text(aes(label=percent(DOS, accuracy=1)), size=3.5, colour="white") +
    scale_fill_viridis_d(end=0.9) +
    scale_y_continuous(expand=c(0,0), breaks=yl, labels=names(yl),
                       sec.axis=dup_axis(breaks=yr, labels=names(yr))) +
    scale_x_continuous(expand=c(0,0)) +
    guides(fill="none") +
    labs(y=NULL)

^{Created on 2025-05-05 with reprex v2.1.1}

[hguthrie]

Not quite. I managed to get a plot that is somewhat close but I cant get the lines to be overlayed on the "filled" parts properly.

library(ggplot2)
library(reshape2)
library(dplyr)
dos_data_long <- melt(dos_data, id.vars = "Energy", variable.name = "Orbital", value.name = "DOS")
head(dos_data_long)

dos_data_filtered <- dos_data_long %>% filter(Orbital != "OPDOS")
dos_data_filtered <- dos_data_filtered %>% filter(Orbital != "Fragment_5")
dos_data_filtered <- dos_data_filtered %>% filter(Orbital != "Fragment_6")
dos_data_filtered <- dos_data_filtered %>% filter(Orbital != "Fragment_7")
dos_data_filtered <- dos_data_filtered %>% filter(Orbital != "Fragment_8")
dos_data_filtered <- dos_data_filtered %>% filter(Orbital != "Fragment_9")
dos_data_filtered <- dos_data_filtered %>% filter(Orbital != "Fragment_10")

#duplicated(dos_data_filtered$Energy)

ggplot(dos_data_filtered, aes(x=Energy, y=DOS, group=Orbital)) +
  geom_area(aes(fill=Orbital), alpha = 0.5) +
  geom_line(color = "black", linewidth=0.5) +
  labs(title="Density of States for Compound [1] Orbitals",
       x = "Energy (eV)",
       y = "DOS") +
  scale_x_continuous() +
  theme_minimal() +
  scale_fill_manual(values = c("TDOS"="gold3", "Ru_s"="wheat2", "Ru_d"="tan3",
                               "Cl_sp"="springgreen4", "Xpy_sp"="slateblue3")) +
  scale_color_manual(values = rep("black", length(unique(dos_data_filtered$Orbital)))) +
  theme(legend.title = element_blank())

It seems as though the lines don't want to line up with the actual data. Although, the DOS for each orbital type does overlap on Energy values but I don't want them stacked. I have seen papers where the DOS look more like this second figure (replied to this) and I would like to get to something similar I just don't know how to get there.
But I am pretty far from that...

[hguthrie]

Something similar to this is what I am aiming for but mine will likely be slightly more complicated because I have alpha and beta spin data. (So I have an additional set of data for each compound and would need to "invert" the second set to see the differences between alpha and beta for each one. I am not sure if that makes sense.

[joels]

For future reference, you can reduce the amount of code needed by doing the filtering in a single filter statement:

dos_data_filtered = dos_data_long %>% 
  filter(!Orbital %in% paste0("Fragment_", 5:10))

[hguthrie]

Do you happen to know why the black lines don't match the edges of the filled data?

[joels]

geom_area() stacks the fill groups by default. geom_line() just plots the lines at their y-values, unstacked, but you can do geom_line(position="stack") to stack them. However, if you just want to add a border line to each area fill, I think geom_area(colour="black") will do it.