library(readxl)
df <- read_excel("D:/Cooliah/Semester 4/Analisis Eksplorasi Data/Pertemuan 3/Data.xlsx", sheet="Sheet2")
df
## # A tibble: 136 × 2
## TPT Waktu
## <chr> <chr>
## 1 5.97 Februari 2022
## 2 5.47 Februari 2022
## 3 6.17 Februari 2022
## 4 4.40 Februari 2022
## 5 4.70 Februari 2022
## 6 4.74 Februari 2022
## 7 3.39 Februari 2022
## 8 4.31 Februari 2022
## 9 4.18 Februari 2022
## 10 8.02 Februari 2022
## # ℹ 126 more rows
df$TPT <- as.double(df$TPT)
library(maps)
library(mapdata)
library(ggplot2)
library(ggrepel)
library(dplyr)
##
## Attaching package: 'dplyr'
## The following objects are masked from 'package:stats':
##
## filter, lag
## The following objects are masked from 'package:base':
##
## intersect, setdiff, setequal, union
ggplot(df, aes(x = as.factor(Waktu), y = TPT)) +
geom_violin(aes(col = Waktu),fill = NA, alpha = 0.2, draw_quantiles = 0.5) +
geom_boxplot(width=.1,fill="orange",outlier.colour=NA)+
stat_summary(fun.y=median,geom="point",fill="blue",shape=21,size=2.5) +
ggtitle("Violin Plot : Tingkat Pengangguran Terbuka di Indonesia
pada Tahun 2022-2023") +
ylab("Tingkat Pengangguran Terbuka") +
xlab("Waktu") +
theme_minimal() +
theme(plot.title = element_text(hjust = 0.5))
## Warning: The `fun.y` argument of `stat_summary()` is deprecated as of ggplot2 3.3.0.
## ℹ Please use the `fun` argument instead.
## This warning is displayed once every 8 hours.
## Call `lifecycle::last_lifecycle_warnings()` to see where this warning was
## generated.
library(readxl)
data <- read_excel("D:/Cooliah/Semester 4/Analisis Eksplorasi Data/Pertemuan 3/Data.xlsx")
data
## # A tibble: 34 × 6
## Provinsi Pulau `Februari 2022` `Agustus 2022` `Februari 2023` `Agustus 2023`
## <chr> <chr> <chr> <chr> <chr> <chr>
## 1 ACEH Suma… 5.97 6.17 5.75 6.03
## 2 SUMATERA… Suma… 5.47 6.16 5.24 5.89
## 3 SUMATERA… Suma… 6.17 6.28 5.9 5.94
## 4 RIAU Suma… 4.40 4.37 4.25 4.23
## 5 JAMBI Suma… 4.70 4.59 4.5 4.53
## 6 SUMATERA… Suma… 4.74 4.63 4.53 4.11
## 7 BENGKULU Suma… 3.39 3.59 3.21 3.42
## 8 LAMPUNG Suma… 4.31 4.52 4.18 4.23
## 9 KEP. BAN… Suma… 4.18 4.77 3.89 4.56
## 10 KEP. RIAU Suma… 8.02 8.23 7.61 6.8
## # ℹ 24 more rows
str(data)
## tibble [34 × 6] (S3: tbl_df/tbl/data.frame)
## $ Provinsi : chr [1:34] "ACEH" "SUMATERA UTARA" "SUMATERA BARAT" "RIAU" ...
## $ Pulau : chr [1:34] "Sumatra" "Sumatra" "Sumatra" "Sumatra" ...
## $ Februari 2022: chr [1:34] "5.97" "5.47" "6.17" "4.40" ...
## $ Agustus 2022 : chr [1:34] "6.17" "6.16" "6.28" "4.37" ...
## $ Februari 2023: chr [1:34] "5.75" "5.24" "5.9" "4.25" ...
## $ Agustus 2023 : chr [1:34] "6.03" "5.89" "5.94" "4.23" ...
data$`Februari 2022`<- as.double(data$`Februari 2022`)
data$`Agustus 2022`<- as.double(data$`Agustus 2022`)
data$`Februari 2023`<- as.double(data$`Februari 2023`)
data$`Agustus 2023`<- as.double(data$`Agustus 2023`)
str(data)
## tibble [34 × 6] (S3: tbl_df/tbl/data.frame)
## $ Provinsi : chr [1:34] "ACEH" "SUMATERA UTARA" "SUMATERA BARAT" "RIAU" ...
## $ Pulau : chr [1:34] "Sumatra" "Sumatra" "Sumatra" "Sumatra" ...
## $ Februari 2022: num [1:34] 5.97 5.47 6.17 4.4 4.7 4.74 3.39 4.31 4.18 8.02 ...
## $ Agustus 2022 : num [1:34] 6.17 6.16 6.28 4.37 4.59 4.63 3.59 4.52 4.77 8.23 ...
## $ Februari 2023: num [1:34] 5.75 5.24 5.9 4.25 4.5 4.53 3.21 4.18 3.89 7.61 ...
## $ Agustus 2023 : num [1:34] 6.03 5.89 5.94 4.23 4.53 4.11 3.42 4.23 4.56 6.8 ...
base::summary(data)
## Provinsi Pulau Februari 2022 Agustus 2022
## Length:34 Length:34 Min. :3.110 Min. :2.340
## Class :character Class :character 1st Qu.:3.985 1st Qu.:4.000
## Mode :character Mode :character Median :4.775 Median :4.685
## Mean :5.123 Mean :4.966
## 3rd Qu.:5.923 3rd Qu.:6.048
## Max. :8.530 Max. :8.310
## Februari 2023 Agustus 2023
## Min. :3.040 Min. :2.270
## 1st Qu.:3.730 1st Qu.:3.487
## Median :4.415 Median :4.320
## Mean :4.806 Mean :4.614
## 3rd Qu.:5.695 3rd Qu.:5.763
## Max. :7.970 Max. :7.520
rata_per_pulau <- data %>%
group_by(Pulau) %>%
summarise(mean_agustus_2023 = mean(`Agustus 2023`, na.rm = TRUE))
rata_per_pulau
## # A tibble: 7 × 2
## Pulau mean_agustus_2023
## <chr> <dbl>
## 1 Jawa 5.86
## 2 Kalimantan 4.56
## 3 Kepulauan Maluku 5.31
## 4 Kepulauan Nusa Tenggara 2.88
## 5 Papua 4.03
## 6 Sulawesi 3.64
## 7 Sumatra 4.97
rata_per_pulau$percentage <- rata_per_pulau$mean_agustus_2023 / sum(rata_per_pulau$mean_agustus_2023) * 100
ggplot(rata_per_pulau, aes(x = "", y = mean_agustus_2023, fill = Pulau)) +
geom_bar(stat = "identity") +
geom_text(aes(label = paste0(round(percentage), "%")), position = position_stack(vjust = 0.5)) +
coord_polar("y", start = 0) +
labs(title = "Komposisi TPT Per Pulau di Indonesia pada Agustus 2023", fill = "Pulau") +
scale_fill_brewer(palette = "Set3") +
theme_void()
library(treemap)
treemap(data,
index = c("Pulau", "Provinsi"),
vSize = "Agustus 2023",
title = "Treemap: Komposisi TPT Per Provinsi pada Agustus 2023",
fontsize.title = 18,
fontsize.labels = 8,
fontcolor.labels = "white"
)
library(ggplot2)
library(dplyr)
data %>%
slice_head(n = 10) %>%
ggplot(aes(y = reorder(Provinsi, `Agustus 2023`), x = `Agustus 2023`)) +
geom_segment(aes(xend = 0, yend = Provinsi), color = "steelblue", size = 2) +
geom_point(color = "lightgreen", size = 5) +
labs(title = "Top 10 Provinsi dengan TPT Tertinggi Pada Agustus 2023", x = "Jumlah", y = "Provinsi")
## Warning: Using `size` aesthetic for lines was deprecated in ggplot2 3.4.0.
## ℹ Please use `linewidth` instead.
## This warning is displayed once every 8 hours.
## Call `lifecycle::last_lifecycle_warnings()` to see where this warning was
## generated.
library(ggplot2)
library(dplyr)
data %>%
slice_tail(n = 10) %>%
ggplot(aes(y = reorder(Provinsi, `Agustus 2023`), x = `Agustus 2023`)) +
geom_segment(aes(xend = 0, yend = Provinsi), color = "steelblue", size = 2) +
geom_point(color = "lightgreen", size = 5) +
labs(title = "10 Provinsi dengan TPT Terendah Pada Agustus 2023", x = "Jumlah", y = "Provinsi")
