How to arrange the order of bars in geom_bar - r

For this week's TidyTuesday challenge on the Erasmus dataset, I was trying to replicate a visualization that seemed cool to me. Yet, I had faced an issue with the ordering of columns in a bar plot and I specifically wanted to ask about a nuance that I could not grasp, although I solved it.
In this first visualization, the coloring of bars was based on their order with regard to their letters.
In the second one, the coloring was based on their order with regard to their total participants.
This all changed thanks to the following code when cleaning data.
mutate(country_name = factor(country_name, levels = rev(country_name)))
Here is also the whole code chunk;
filter (receiving_country_code == "TR") %>%
filter (sending_country_code != receiving_country_code) %>%
group_by(sending_country_code) %>%
summarize(total = sum(participants)) %>%
arrange(desc(total)) %>%
filter(total>9) %>%
left_join(country_codes, by = c("sending_country_code" = "iso2c")) %>%
mutate(country_name = case_when(
sending_country_code == "UK" ~ "United Kingdom",
sending_country_code == "EL" ~ "Greece",
TRUE ~ country_name
)) %>%
mutate(country_name = factor(country_name, levels = rev(country_name)))
p1 <- ggplot(data = top_incoming,
aes(x = country_name, y = total, fill = country_name)) +
geom_bar(width = 0.3, stat = "identity", show.legend = FALSE) +
geom_text(aes(x = country_name, y = 0, label = paste0(country_name, " - ", total)),
hjust = 1.05, size = 2, colour = rep(c("#173f5f", "#3caea3"), 10)) +
coord_polar(theta = "y", start = 0) +
ylim(c(0, 40)) +
scale_fill_manual(values = rep(c("#3caea3", "#173f5f"), 10)) +
labs(title = "Incoming Students",
subtitle = "2015-2019") +
theme_void() +
theme (plot.title = element_text(hjust = 0.5, family = "subtitle", size = 14),
plot.subtitle = element_text(hjust = 0.5, family = "subtitle", size = 12),
plot.caption = element_text(family = "caption"))
This all changed thanks to the following code when cleaning data.
Although I figured out what this code's function is, I could not understand its underlying logic, particularly what levels = rev(country_name) means considering that the order of the columns was primarily based on character level and, at best, I think, the  rev() function should have reversed it instead of changing the order to be based on participants. 

Related

Text color with geom_label_repel

Not specific to any particular piece of code, is there a relatively straightforward way to change the color of the text in a geom_label_repel box?
Specifically, I have code that produces the below chart
The percentage in the label box is the percent change in 7-day moving average for the most recent week over the week prior. I'd simply like to color the text red when the value is positive and green when it is negative.
The dataframe for this chart can be copied from here.
The plot code is
#endpoint layer
BaseEndpoints <- smDailyBaseData %>% filter(Base %in% AFMCbases) %>%
group_by(Base) %>%
filter(DaysSince == max(DaysSince)) %>%
select(Base, abbv, DaysSince, newRate,label) %>%
ungroup()
ZoomEndpoints <- BaseEndpoints %>% filter(Base != 'Edwards') %>%
mutate(zoom = TRUE)
CAEndPoint <- BaseEndpoints %>% filter(Base == 'Edwards') %>%
mutate(zoom = FALSE)
ZoomEndpoints <- rbind(ZoomEndpoints, CAEndPoint)
BasePlot <- smDailyBaseData %>% filter(Base %in% AFMCbases) %>%
ggplot(mapping = aes(x = as.numeric(DaysSince), y = newRate)) +
geom_line(aes(color=abbv),show.legend = FALSE) +
scale_color_ucscgb() +
geom_point(data = BaseEndpoints,size = 1.5,shape = 21,
aes(color = abbv,fill = abbv), show.legend = FALSE) +
geom_label_repel(data=ZoomEndpoints, aes(label=label), show.legend = FALSE,
vjust = 0, xlim=c(105,200), size=3, direction='y') +
labs(x = "Days Since First Confirmed Case",
y = "% Local Population Infected Daily") +
theme(plot.title = element_text(size = rel(1), face = "bold"),
plot.subtitle = element_text(size = rel(0.7)),
plot.caption = element_text(size = rel(1))) +
facet_zoom(xlim = c(50,120), ylim=c(0,0.011),zoom.data=zoom)
print(BasePlot)
Yes, it's as simple as this:
library(ggplot2)
df <- data.frame(x = c(-1, -1, 1, 1), y = c(-1, 1, 1, -1), value = c(-2, -1, 1, 2))
ggplot(df, aes(x, y)) +
geom_point(size = 3) +
ggrepel::geom_label_repel(aes(label = value, colour = factor(sign(value)))) +
lims(x = c(-100, 100), y = c(-100, 100)) +
scale_colour_manual(values = c("red", "forestgreen"))
EDIT
Now we have a more concrete example, I can see the problem more clearly. There are workarounds such as using ggnewscale or a hand-crafted solution such as Ian Campbell's thorough example. Personally, I would just note that you haven't used the fill scale yet, and this looks pretty good to my eye:
Here's a bit of a hacky solution since you can't have two scale_color_*'s at the same time:
The approach centers on manually assigning the color outside of aes in the geom_label_repel call. Adding one to the grepl result that searches for the minus sign in the label allows you to subset the two colors. You need two colors for each label, I assume for the box and for the text, so I used rep.
smDailyBaseData %>%
ggplot(mapping = aes(x = as.numeric(DaysSince), y = newRate)) +
geom_line(aes(color=abbv),show.legend = FALSE) +
scale_color_ucscgb() +
geom_point(data = BaseEndpoints,size = 1.5,shape = 21,
aes(color = abbv,fill = abbv), show.legend = FALSE) +
geom_label_repel(data=ZoomEndpoints, aes(label=label),
color = rep(c("green","red")[1+grepl("\\-\\d",as.factor(ZoomEndpoints$label))],times = 2),
show.legend = FALSE, vjust = 0, xlim=c(105,200),
size=3, direction='y') +
labs(x = "Days Since First Confirmed Case",
y = "% Local Population Infected Daily") +
theme(plot.title = element_text(size = rel(1), face = "bold"),
plot.subtitle = element_text(size = rel(0.7)),
plot.caption = element_text(size = rel(1))) +
facet_zoom(xlim = c(50,120), ylim=c(0,0.011),zoom.data=zoom)
Data Setup
#source("https://pastebin.com/raw/Vn2abQ4a")
BaseEndpoints <- smDailyBaseData %>%
group_by(Base) %>%
dplyr::filter(DaysSince == max(DaysSince)) %>%
dplyr::select(Base, abbv, DaysSince, newRate,label) %>%
ungroup()
ZoomEndpoints <- BaseEndpoints %>% filter(Base != 'Edwards') %>%
mutate(zoom = TRUE)
CAEndPoint <- BaseEndpoints %>% filter(Base == 'Edwards') %>%
mutate(zoom = FALSE)
ZoomEndpoints <- rbind(ZoomEndpoints, CAEndPoint)

Animated sorted bar chart with bars overtaking each other

Edit: keyword is 'bar chart race'
How would you go at reproducing this chart from Jaime Albella in R ?
See the animation on visualcapitalist.com or on twitter (giving several references in case one breaks).
I'm tagging this as ggplot2 and gganimate but anything that can be produced from R is relevant.
data (thanks to https://github.com/datasets/gdp )
gdp <- read.csv("https://raw.github.com/datasets/gdp/master/data/gdp.csv")
# remove irrelevant aggregated values
words <- scan(
text="world income only total dividend asia euro america africa oecd",
what= character())
pattern <- paste0("(",words,")",collapse="|")
gdp <- subset(gdp, !grepl(pattern, Country.Name , ignore.case = TRUE))
Edit:
Another cool example from John Murdoch :
Most populous cities from 1500 to 2018
Edit: added spline interpolation for smoother transitions, without making rank changes happen too fast. Code at bottom.
I've adapted an answer of mine to a related question. I like to use geom_tile for animated bars, since it allows you to slide positions.
I worked on this prior to your addition of data, but as it happens, the gapminder data I used is closely related.
library(tidyverse)
library(gganimate)
library(gapminder)
theme_set(theme_classic())
gap <- gapminder %>%
filter(continent == "Asia") %>%
group_by(year) %>%
# The * 1 makes it possible to have non-integer ranks while sliding
mutate(rank = min_rank(-gdpPercap) * 1) %>%
ungroup()
p <- ggplot(gap, aes(rank, group = country,
fill = as.factor(country), color = as.factor(country))) +
geom_tile(aes(y = gdpPercap/2,
height = gdpPercap,
width = 0.9), alpha = 0.8, color = NA) +
# text in x-axis (requires clip = "off" in coord_*)
# paste(country, " ") is a hack to make pretty spacing, since hjust > 1
# leads to weird artifacts in text spacing.
geom_text(aes(y = 0, label = paste(country, " ")), vjust = 0.2, hjust = 1) +
coord_flip(clip = "off", expand = FALSE) +
scale_y_continuous(labels = scales::comma) +
scale_x_reverse() +
guides(color = FALSE, fill = FALSE) +
labs(title='{closest_state}', x = "", y = "GFP per capita") +
theme(plot.title = element_text(hjust = 0, size = 22),
axis.ticks.y = element_blank(), # These relate to the axes post-flip
axis.text.y = element_blank(), # These relate to the axes post-flip
plot.margin = margin(1,1,1,4, "cm")) +
transition_states(year, transition_length = 4, state_length = 1) +
ease_aes('cubic-in-out')
animate(p, fps = 25, duration = 20, width = 800, height = 600)
For the smoother version at the top, we can add a step to interpolate the data further before the plotting step. It can be useful to interpolate twice, once at rough granularity to determine the ranking, and another time for finer detail. If the ranking is calculated too finely, the bars will swap position too quickly.
gap_smoother <- gapminder %>%
filter(continent == "Asia") %>%
group_by(country) %>%
# Do somewhat rough interpolation for ranking
# (Otherwise the ranking shifts unpleasantly fast.)
complete(year = full_seq(year, 1)) %>%
mutate(gdpPercap = spline(x = year, y = gdpPercap, xout = year)$y) %>%
group_by(year) %>%
mutate(rank = min_rank(-gdpPercap) * 1) %>%
ungroup() %>%
# Then interpolate further to quarter years for fast number ticking.
# Interpolate the ranks calculated earlier.
group_by(country) %>%
complete(year = full_seq(year, .5)) %>%
mutate(gdpPercap = spline(x = year, y = gdpPercap, xout = year)$y) %>%
# "approx" below for linear interpolation. "spline" has a bouncy effect.
mutate(rank = approx(x = year, y = rank, xout = year)$y) %>%
ungroup() %>%
arrange(country,year)
Then the plot uses a few modified lines, otherwise the same:
p <- ggplot(gap_smoother, ...
# This line for the numbers that tick up
geom_text(aes(y = gdpPercap,
label = scales::comma(gdpPercap)), hjust = 0, nudge_y = 300 ) +
...
labs(title='{closest_state %>% as.numeric %>% floor}',
x = "", y = "GFP per capita") +
...
transition_states(year, transition_length = 1, state_length = 0) +
enter_grow() +
exit_shrink() +
ease_aes('linear')
animate(p, fps = 20, duration = 5, width = 400, height = 600, end_pause = 10)
This is what I came up with, so far, based in good part on #Jon's answer.
p <- gdp %>%
# build rank, labels and relative values
group_by(Year) %>%
mutate(Rank = rank(-Value),
Value_rel = Value/Value[Rank==1],
Value_lbl = paste0(" ",round(Value/1e9))) %>%
group_by(Country.Name) %>%
# keep top 10
filter(Rank <= 10) %>%
# plot
ggplot(aes(-Rank,Value_rel, fill = Country.Name)) +
geom_col(width = 0.8, position="identity") +
coord_flip() +
geom_text(aes(-Rank,y=0,label = Country.Name,hjust=0)) + #country label
geom_text(aes(-Rank,y=Value_rel,label = Value_lbl, hjust=0)) + # value label
theme_minimal() +
theme(legend.position = "none",axis.title = element_blank()) +
# animate along Year
transition_states(Year,4,1)
animate(p, 100, fps = 25, duration = 20, width = 800, height = 600)
I might come back to improve it.
The moving grid could be simulated by removing the actual grid and having geom_segment lines moving and fading out thanks to an alpha parameter changing when it approaches 100 billion.
To have labels changing values between years (which gives a nice feeling of urgency in the original chart) I think we have no choice but multiplying the rows while interpolating labels, we'll need to interpolate Rank too.
Then with a few minor cosmetic changes we should be pretty close.
This is what I came up, I just use Jon and Moody code as a template and make few changes.
library(tidyverse)
library(gganimate)
library(gapminder)
theme_set(theme_classic())
gdp <- read.csv("https://raw.github.com/datasets/gdp/master/data/gdp.csv")
words <- scan(
text="world income only total dividend asia euro america africa oecd",
what= character())
pattern <- paste0("(",words,")",collapse="|")
gdp <- subset(gdp, !grepl(pattern, Country.Name , ignore.case = TRUE))
colnames(gdp) <- gsub("Country.Name", "country", colnames(gdp))
colnames(gdp) <- gsub("Country.Code", "code", colnames(gdp))
colnames(gdp) <- gsub("Value", "value", colnames(gdp))
colnames(gdp) <- gsub("Year", "year", colnames(gdp))
gdp$value <- round(gdp$value/1e9)
gap <- gdp %>%
group_by(year) %>%
# The * 1 makes it possible to have non-integer ranks while sliding
mutate(rank = min_rank(-value) * 1,
Value_rel = value/value[rank==1],
Value_lbl = paste0(" ",value)) %>%
filter(rank <=10) %>%
ungroup()
p <- ggplot(gap, aes(rank, group = country,
fill = as.factor(country), color = as.factor(country))) +
geom_tile(aes(y = value/2,
height = value,
width = 0.9), alpha = 0.8, color = NA) +
geom_text(aes(y = 0, label = paste(country, " ")), vjust = 0.2, hjust = 1) +
geom_text(aes(y=value,label = Value_lbl, hjust=0)) +
coord_flip(clip = "off", expand = FALSE) +
scale_y_continuous(labels = scales::comma) +
scale_x_reverse() +
guides(color = FALSE, fill = FALSE) +
labs(title='{closest_state}', x = "", y = "GDP in billion USD",
caption = "Sources: World Bank | Plot generated by Nitish K. Mishra #nitishimtech") +
theme(plot.title = element_text(hjust = 0, size = 22),
axis.ticks.y = element_blank(), # These relate to the axes post-flip
axis.text.y = element_blank(), # These relate to the axes post-flip
plot.margin = margin(1,1,1,4, "cm")) +
transition_states(year, transition_length = 4, state_length = 1) +
ease_aes('cubic-in-out')
animate(p, 200, fps = 10, duration = 40, width = 800, height = 600, renderer = gifski_renderer("gganim.gif"))
Here I am using duration 40 second, which is slow. You can change duration and make it faster or slower as you needed.

R - How can I add a bivariate legend to my ggplot2 chart?

I'm trying to add a bivariate legend to my ggplot2 chart but I don't know whether (a) this is possible through some guides options and (b) how to achieve it.
The only way I've managed to produce something close to the desired outcome was by specifically creating a new chart which resembles a legend (named p.legend below) and inserting it, via the cowplot package, somewhere in the original chart (named p.chart below). But surely there must be a better way than this, given that this approach requires creating the legend in the first place and fiddling with its size/location to fit it in the original chart.
Here's code for a dummy example of my approach:
library(tidyverse)
# Create Dummy Data #
set.seed(876)
n <- 2
df <- expand.grid(Area = LETTERS[1:n],
Period = c("Summer", "Winter"),
stringsAsFactors = FALSE) %>%
mutate(Objective = runif(2 * n, min = 0, max = 2),
Performance = runif(2 * n) * Objective) %>%
gather(Type, Value, Objective:Performance)
# Original chart without legend #
p.chart <- df %>%
ggplot(., aes(x = Area)) +
geom_col(data = . %>% filter(Type == "Objective"),
aes(y = Value, fill = Period),
position = "dodge", width = 0.7, alpha = 0.6) +
geom_col(data = . %>% filter(Type == "Performance"),
aes(y = Value, fill = Period),
position = "dodge", width = 0.7) +
scale_fill_manual(values = c("Summer" = "#ff7f00", "Winter" = "#1f78b4"), guide = FALSE) +
theme_minimal() +
theme(panel.grid.major.x = element_blank(),
panel.grid.minor.y = element_blank())
# Create a chart resembling a legend #
p.legend <- expand.grid(Period = c("Summer", "Winter"),
Type = c("Objective", "Performance"),
stringsAsFactors = FALSE) %>%
ggplot(., aes(x = Period, y = factor(Type, levels = c("Performance", "Objective")),
fill = Period, alpha = Type)) +
geom_tile() +
scale_fill_manual(values = c("Summer" = "#ff7f00", "Winter" = "#1f78b4"), guide = FALSE) +
scale_alpha_manual(values = c("Objective" = 0.7, "Performance" = 1), guide = FALSE) +
ggtitle("Legend") +
theme_minimal() +
theme(plot.title = element_text(hjust = 0.5),
rect = element_rect(fill = "transparent"),
axis.title = element_blank(),
panel.grid.major = element_blank())
# Add legend to original chart #
p.final <- cowplot::ggdraw() +
cowplot::draw_plot(plot = p.chart) +
cowplot::draw_plot(plot = p.legend, x = 0.5, y = 0.65, width = 0.4, height = 0.28, scale = 0.7)
# Save chart #
cowplot::ggsave("Bivariate Legend.png", p.final, width = 8, height = 6, dpi = 500)
... and the resulting chart:
Is there an easier way of doing this?
This might work at some point, but right now the colorbox seems to ignore all breaks, names and labels (#ClausWilke?). Probably because the multiscales package is in really early stages.
Posting since it might work when future readers are here.
library(multiscales)
df %>%
mutate(
period = as.numeric(factor(Period)),
type = as.numeric(factor(Type))
) %>%
ggplot(., aes(x = Area, y = Value, fill = zip(period, type), group = interaction(Area, Period))) +
geom_col(width = 0.7, position = 'dodge') +
bivariate_scale(
"fill",
pal_hue_sat(c(0.07, 0.6), c(0.4, 0.8)),
guide = guide_colorbox(
nbin = 2,
name = c("Period", "Type"), #ignored
breaks = list(1:2, 1:2), #ignored
labels = list(levels(.$Period), levels(.$Type)) #ignored
)

Ribbon chart in R

I search in R implementation (may be html widget on java script) a stacked bar chart in ribbon style, which allows you to see the rating change for each category in the dynamics.
It's look like ribbon chart in power bi desktop
Search rseek.org gave no results.
First off: Not a fan of that ribbon-styled stacked bar chart at all; while colourful and stylish, it's difficult to synthesise the relevant information. But that's just my opinion.
You could try building a similar plot in ggplot2 using geom_ribbon. See below for a minimal example:
# Sample data
set.seed(2017);
one <- sample(5:15, 10);
two <- rev(one);
df <- cbind.data.frame(
x = rep(1:10, 2),
y = c(one, two),
l = c(one - 1, two - 1),
h = c(one + 1, two + 1),
id = rep(c("one", "two"), each = 10));
require(ggplot2);
ggplot(df, aes(x = x, y = y)) +
geom_ribbon(aes(ymin = l, ymax = h, fill = id), alpha = 0.4) +
scale_fill_manual(values = c("#E69F00", "#56B4E9"));
If you need interactivity, you could wrap it inside plotly::ggplotly.
Using ggsankey package.
In the following you can make use of smooth argument geom_sankey_bump to control the look/feel of the chart as in ribbon chart of Power BI.
df <- data.frame (model = c("A","B","C","D","E","F","G","H","I","J","A","B","C","D","E","F","G","H","I","J","A","B","C","D","E","F","G","H","I","J","A","B","C","D","E","F","G","H","I","J"),
Year = c(2015,2015,2015,2015,2015,2015,2015,2015,2015,2015,2016,2016,2016,2016,2016,2016,2016,2016,2016,2016,2017,2017,2017,2017,2017,2017,2017,2017,2017,2017,2018,2018,2018,2018,2018,2018,2018,2018,2018,2018),
sales = c(450,678,456,344,984,456,234,244,655,789,234,567,234,567,232,900,1005,1900,450,345,567,235,456,345,144,333,555,777,111,444,222,223,445,776,331,788,980,1003,456,434))
#install.packages("remotes")
#remotes::install_github("davidsjoberg/ggsankey")
library(ggsankey)
library(tidyverse)
ggplot(df, aes(x = Year,
node = model,
fill = model,
value = sales)) +
geom_sankey_bump(space = 0, type = "alluvial", color = "transparent", smooth = 15) +
scale_fill_viridis_d(option = "A", alpha = .8) +
theme_sankey_bump(base_size = 16) +
labs(x = NULL,
y = "Sales ($ ths)",
fill = "Model",
color = NULL) +
theme(legend.position = "bottom") +
labs(title = "Sales per model per year")
On suggestion in comments, I tried replicating some of the features of power BI chart.
# Prepare some data
set.seed(1)
df <- data.frame(
occupation = rep(c("Clerical", "Management", "Manual", "Professional", "Skilled"), 12),
Month = factor(rep(month.abb, 5), levels = month.abb, ordered = TRUE),
Sales = sample(200:1000, 60, replace = TRUE)
)
df %>%
group_by(Month) %>%
mutate(Max = sum(Sales)) %>%
ungroup() %>%
mutate(Max = max(Sales)) %>%
ggplot(aes(x = Month,
node = occupation,
fill = occupation,
value = Sales)) +
geom_col(aes(x = Month, y = Max/1.2),
alpha = 0.5,
fill = 'grey',
width = 0.4) +
geom_sankey_bump(space = 15,
type = "alluvial",
color = "transparent",
smooth = 8,
alpha = 0.8) +
scale_fill_brewer(palette = "Set3") +
theme_minimal() +
labs(x = NULL,
y = "Sales ($ ths)",
fill = "Occupation",
color = NULL) +
theme(legend.position = "top") +
labs(title = "Sales per occupation per month")
Created on 2022-07-07 by the reprex package (v2.0.1)
You may find your answers with ggalluvial package.
https://cran.r-project.org/web/packages/ggalluvial/vignettes/ggalluvial.html

Reordering legend items in ggplot from two different datasets and layers

I'm combining two layers in ggplot that were created from two different data sets and want to control the order in which the legend appears.
With example data and code:
base <-
data.frame(idea_num = c(1, 2),
value = c(-50, 90),
it_cost = c(30, 10))
group <-
data.frame(idea_num = c(1, 1, 2, 2),
group = c("a", "b", "a", "b"),
is_primary = c(TRUE, FALSE, FALSE, TRUE),
group_value = c(-40, -10, 20, 70))
base %>%
left_join(group) %>%
arrange(desc(value)) %>%
mutate(idea_num = idea_num %>% factor(levels = unique(idea_num)),
is_primary = is_primary %>% factor(levels = c("TRUE", "FALSE"))) %>%
ggplot(aes(x = idea_num, y = group_value, fill = is_primary)) +
geom_bar(stat = "identity") +
geom_bar(data = base %>%
arrange(desc(value)) %>%
mutate(idea_num = idea_num %>% factor(levels = unique(idea_num))),
aes(x = idea_num, y = it_cost, alpha = 0.1, fill = "it_cost"),
stat = "identity") +
scale_fill_manual(name = "Group", labels = c("TRUE" = "Primary", "FALSE" = "Secondary", "it_cost" = "IT Cost"),
values = c("TRUE" = "blue", "FALSE" = "red", "it_cost" = "black")) +
scale_alpha(guide = "none") +
theme(legend.position = "bottom")
I get a figure
but I'd like the legend to appear in the order of Primary, Secondary, IT Cost.
Were all of the numbers I'm trying to plot part of the same grand number, I could easily melt the dataframe and sum everything; however, the values from the group$group_value need to be displayed separate from base$it_cost.
If I plot only the values from teh first layer, i.e.,
base %>%
left_join(group) %>%
arrange(desc(value)) %>%
mutate(idea_num = idea_num %>% factor(levels = unique(idea_num)),
is_primary = is_primary %>% factor(levels = c("TRUE", "FALSE"))) %>%
ggplot(aes(x = idea_num, y = group_value, fill = is_primary)) +
geom_bar(stat = "identity") +
scale_fill_manual(name = "Group", labels = c("TRUE" = "Primary", "FALSE" = "Secondary"),
values = c("TRUE" = "blue", "FALSE" = "red")) +
theme(legend.position = "bottom")
I get a figure I expect
How can I add the second layer and adjust the ordering of the legend boxes? I do not believe that this question or this question are entirely relevant to mine as the former is dealing with levels of a factor and the latter deals with ordering of multiple legends.
Can I do what I'd like to do? Is there a better way of constructing this plot?
use scale_fill_manual(..., limit=, ...):
... +
scale_fill_manual(name = "Group",
labels = c("TRUE" = "Primary", "FALSE" = "Secondary", "it_cost" = "IT Cost"),
limits = c("TRUE", "FALSE", "it_cost"),
values = c("TRUE" = "blue", "FALSE" = "red", "it_cost" = "black")) +
...
This gives:
That said, I think you may want to consider a few different approaches:
A: why do you create your data in such a complex way, ending up multiple observations of IT Costs for the same idea number? I don't know your data, you may well have your reasons, but a simple dataset along the lines:
idea_num value type
1 1 -40 Primary
2 1 -10 Secondary
3 2 20 Secondary
4 2 70 Primary
5 1 -50 IT Cost
6 2 90 IT Cost
would simplify the things quite a bit.
B: Why do you want to stack/overplot these two separate barplots? I would do position="dodge" instead to have separate bars.
df2 <- base %>%
left_join(group) %>%
mutate(is_primary=paste0("pri_", is_primary+0)) %>%
spread(is_primary, group_value) %>%
gather(yvar, y, it_cost, pri_0, pri_1)
df2$yvar <- factor(df2$yvar, levels=c("pri_0", "pri_1", "it_cost"),
labels=c("Primary", "Secondary", "IT Cost"))
df2$idea_num <- factor(df2$idea_num, levels=c(2, 1))
ggplot(df2, aes(idea_num, y, fill=yvar)) +
geom_bar(stat="identity") +
scale_fill_manual("Group", values=c("blue", "red", "black")) +
scale_alpha(guide = "none") +
theme(legend.position = "bottom")

Resources