Histogram with density curves of other data - r

Question

I want to recreate but I have trouble to fit the density curves to the plot.
MWE ( of what I achieved so far. Data in the tibble are just sample data) :
tibble(home = sample(1:10,90, replace = T), away = sample(1:10,90, replace = T)) %>%
gather(key=Type, value=Value) %>%
ggplot(aes(x=Value,fill=Type)) +
geom_histogram(position="dodge")
UPDATE after answer by #Kota Mori
I adjusted the answer given by Kota Mori to get the following which results in an error. Before I start lets have a look at the datasets I want to use for the graph :
#Both Goals variables of this dataframe should be used for the histogram
actual
# A tibble: 90 x 7
season matchday club_name_home club_name_away goals_team_home goals_team_away sumgoals
<dbl> <dbl> <chr> <chr> <dbl> <dbl> <dbl>
1 1819 21 ETuS Haltern TuS 05 Sinsen II 2 2 4
2 1819 21 VfL Ramsdorf Westfalia Gemen II 2 0 2
3 1819 21 FC RW Dorsten SV Altendorf-Ulfkotte 8 4 12
4 1819 21 SuS Hervest-Dorsten 1. SC BW Wulfen 0 0 0
5 1819 21 SV Lembeck SC Reken II 1 1 2
6 1819 21 RC Borken-Hoxfeld TSV Raesfeld 3 1 4
7 1819 21 TuS Velen Fenerbahce I. Marl 5 2 7
8 1819 21 BVH Dorsten SC Marl-Hamm 2 0 2
9 1819 21 1. SC BW Wulfen FC RW Dorsten 3 0 3
10 1819 21 BVH Dorsten SV Altendorf-Ulfkotte 2 0 2
# ... with 80 more rows
#Both Goals variables of this dataframe should be used for the density lines
poisson
# A tibble: 90 x 6
season matchday club_name_home club_name_away Goals_team_home Goals_team_away
<dbl> <dbl> <chr> <chr> <dbl> <dbl>
1 1819 21 ETuS Haltern TuS 05 Sinsen II 2 2
2 1819 21 VfL Ramsdorf Westfalia Gemen II 3 0
3 1819 21 FC RW Dorsten SV Altendorf-Ulfkotte 2 0
4 1819 21 SuS Hervest-Dorsten 1. SC BW Wulfen 0 4
5 1819 21 SV Lembeck SC Reken II 2 1
6 1819 21 RC Borken-Hoxfeld TSV Raesfeld 2 1
7 1819 21 TuS Velen Fenerbahce I. Marl 2 1
8 1819 21 BVH Dorsten SC Marl-Hamm 3 1
9 1819 21 1. SC BW Wulfen FC RW Dorsten 2 0
10 1819 21 BVH Dorsten SV Altendorf-Ulfkotte 2 1
# ... with 80 more rows
So I adjusted the answer by Kota Mori to end up with the following code :
simyears = 1819
actual <- read_rds(here::here(paste0("/data/database_match_results_",sim_years,".rds")))%>%
filter(between(matchday, 21, max(database_season$matchday)))
poisson <- missinggames
data <- rbind(data.frame(type="home", value=actual$goals_team_home, stringsAsFactors=FALSE),
data.frame(type="away", value=actual$goals_team_home, stringsAsFactors=FALSE))
estimate <- group_by(poisson %>% select(Goals_team_home,Goals_team_away), type) %>% summarize(mu=mean(value))
dens <- expand.grid(value=0:max(data$value), type=c("away", "home"),
stringsAsFactors=FALSE) %>%
inner_join(estimate) %>%
mutate(density=dpois(value, mu))
prop <- group_by(data, type, value) %>% summarize(count=n()) %>%
group_by(type) %>% mutate(prop=count/sum(count))
tmp_actual <- left_join(dens, prop) %>% replace_na(list(prop=0, count=0))
ggplot(tmp_actual, aes(x=value, weight=prop, fill=type)) +
geom_bar(position="dodge") +
geom_line(aes(value, density, color=type, weight=NULL))
Which results in the following error : 'Error: Mapping should be created with aes() or aes_().'

Answer 1

I think you need to calculate the poisson parameters on your own, which turns out to be as easy as calculating the sample mean for each type.
The following code generates a graph similar to the example.
library(dplyr)
library(ggplot2)
data <- rbind(data.frame(type="home", value=rpois(90, 2.5), stringsAsFactors=FALSE),
data.frame(type="away", value=rpois(90, 1.5), stringsAsFactors=FALSE))
estimate <- group_by(data, type) %>% summarize(mu=mean(value))
dens <- expand.grid(value=0:max(data$value), type=c("away", "home"),
stringsAsFactors=FALSE) %>%
inner_join(estimate) %>%
mutate(density=dpois(value, mu))
prop <- group_by(data, type, value) %>% summarize(count=n()) %>%
group_by(type) %>% mutate(prop=count/sum(count))
tmp <- left_join(dens, prop) %>% replace_na(list(prop=0, count=0))
ggplot(tmp, aes(x=value, weight=prop, fill=type)) +
geom_bar(position="dodge") +
geom_line(aes(value, density, color=type, weight=NULL))