Dataset: Financial Inclusion in Africa
0.1 Set up
knitr::opts_chunk$set(echo = TRUE,
message=FALSE,warning=FALSE,
fig.width = 8, fig.height = 6)
0.2 Install and load the packages required
### create a vector of packages to be installed
pkgs <- c("tidyverse","DT","lubridate","patchwork")
### Check if there are packages you want to load, that are not already installed.
miss_pkgs <- pkgs[!pkgs %in% installed.packages()[,1]]
### Installing the missing packages
if(length(miss_pkgs)>0){
install.packages(miss_pkgs)
}
### Loading all the packages
invisible(lapply(pkgs,library,character.only=TRUE))
### Remove the objects that are no longer required
rm(miss_pkgs)
rm(pkgs)
# Set the theme
# mir_theme <- theme(plot.title = element_text(size = 12, #18
# #family = "Source Sans Pro Semibold",
# face = "italic", hjust = 0.5),
# axis.line = element_line(color = "black", size = 1),
# axis.title = element_text(size = 16),
# axis.text = element_text(size = 14),
# panel.background = element_rect(fill = NA),
# plot.caption = element_text(size = 14),
# legend.title = element_blank(),
# legend.position = "bottom")
mir_theme <- theme(plot.title = element_text(size = 12, #18
#family = "Source Sans Pro Semibold",
face = "italic", hjust = 0.5),
axis.line = element_line(color = "black", size = 1),
axis.title = element_text(size = 10),
axis.text = element_text(size = 9),
panel.background = element_rect(fill = NA),
plot.caption = element_text(size = 8),
legend.title = element_blank(),
legend.position = "bottom")
1. Read in the dataset
df <- read_csv("Train_v2.csv")
2. Convert character variables to factor variables
## 2.1. bank_account
df <- df %>%
mutate(bank_account = fct_relevel(bank_account, "No", "Yes"))
## 2.2. location_type
df <- df %>%
mutate(location_type = fct_relevel(location_type, "Rural" , "Urban"))
## 2.3. cellphone_access
df <- df %>%
mutate(cellphone_access = fct_relevel(cellphone_access, "No", "Yes"))
## 2.4. gender_of_respondent
df <- df %>%
mutate(gender_of_respondent = fct_relevel(gender_of_respondent, "Female" ,"Male"))
## 2.5. relationship_with_head
df <- df %>%
mutate(relationship_with_head = fct_relevel(relationship_with_head,
"Child","Spouse", "Parent", "Head of Household",
"Other relative", "Other non-relatives" ))
## 2.6. marital_status
df <- df %>%
mutate(marital_status = fct_relevel(marital_status,
"Single/Never Married","Divorced/Seperated","Widowed",
"Married/Living together","Dont know"))
## 2.7. education_level
df <- df %>%
mutate(education_level = fct_relevel(education_level,
"No formal education", "Primary education", "Secondary education",
"Tertiary education", "Vocational/Specialised training","Other/Dont know/RTA"))
## 2.8. job_type
3. Generate 2 new variables i.e monthly income and date they registered for xyz
## 3.1 monthly income
set.seed(2020)
income_values <- sample(c(5000 : 150000), nrow(df))
df <- df %>%
mutate(income = income_values,
income = ifelse(job_type == "No Income", NA, income))
## 3.2 date of registering for xyz
date_vec <- sample(seq(as.Date('2014/01/01'), as.Date('2014/12/31'), by="day"), nrow(df), replace = T)
df <- df %>%
mutate(date = date_vec)
## Generate the month
df <- df %>%
mutate(month = month(date, abbr = T, label = T))
4. Main graphs
4.1 Bar graph
4.1.1 Single categorical variable
Distribution of cellphone access
## First generate a table
tab1 <- df %>%
group_by(cellphone_access) %>%
count() %>%
ungroup() %>%
mutate(perc = round(n/sum(n) *100, 0))
## Plot a graph
plot1 <- ggplot(data = tab1, aes(x = cellphone_access, y = perc))+
geom_bar(stat = "identity", fill = "maroon", width = 0.7)+
geom_text(aes(label = perc), size = 5, hjust = 0.5, vjust = -0.25)+
mir_theme +
labs(title = "Distribution of cellphone access",
x = "",y = "Percentage")+
#caption = "Twitter:@Shel_Kariuki")+
ylim(c(0,100))
plot1
#rstudio_blue <- "#4AA4DE"
Distribution of sample by country
## table
tab2 <- df %>%
group_by(country) %>%
count() %>%
ungroup() %>%
mutate(perc = round(n / sum(n) *100,1))
## graph
plot2 <- ggplot(data = tab2, aes(x = country, y=perc)) +
geom_bar(stat = "identity", fill = "maroon", width = 0.7) +
geom_text(aes(label = perc), size = 4 , hjust = 0.5 , vjust = -0.25)+
mir_theme +
labs(title = "Distribution of sample by country",
x = "", y = "Percentage",
caption = "Twitter: @Shel_Kariuki")
plot2
4.1.2 Two categorical variables (main variable and grouping variable)
Distribution of bank account availability by country
## table
tab3 <- df %>%
group_by(country, bank_account) %>%
count() %>%
ungroup() %>%
group_by(bank_account) %>%
mutate(perc = round(n / sum(n) *100,1))
## graph
plot3 <- ggplot(data = tab3, aes(x = country, y=perc, fill = bank_account)) +
geom_bar(stat = "identity", width = 0.7, position = "dodge") +
geom_text(aes(label = perc), size = 4 , hjust = 0.5 , vjust = -0.25,
position = position_dodge(width = 0.7))+
mir_theme +
scale_fill_manual(values = c("grey", "maroon"))+
labs(title = "Distribution of bank account availability by country",
x = "", y = "Percentage",
caption = "Twitter: @Shel_Kariuki")+
ylim(c(0, 50))
plot3
4.2 Line graph
4.2.1 Single categorical variable
Distribution of registrations by month
## table
tab4 <- df %>%
group_by(month) %>%
count() %>%
ungroup() %>%
mutate(perc = round(n/sum(n)*100,1))
## graph
plot4 <- ggplot(data = tab4, aes(x = month, y = perc, group = 1)) +
geom_point()+
geom_line(stat = "identity", color = "maroon", size = 1)+
geom_text(aes(label = perc), size = 5, hjust = 0.5, vjust = -0.25)+
mir_theme +
labs(title = "Distribution of registrations by month",
x = "", y = "Percentage",
caption = "Twitter: @Shel_Kariuki")+
ylim(c(7.5,9))
plot4
Average age of respondents registering per month
## table
tab5 <- df %>%
group_by(month) %>%
summarise(avg_age = round(mean(age_of_respondent, na.rm = T),1))
## plot
plot5 <- ggplot(data = tab5, aes(x = month, y = avg_age, group = 1))+
geom_point()+
geom_line(stat = "identity", size = 1, color = "maroon", linetype = "solid")+
geom_text(aes(label = avg_age), vjust = 0, hjust = 1)+
mir_theme +
labs(title = "Average age of respondents registering per month",
x = "", y = "Average Age",
caption = "Twitter: @Shel_Kariuki")+
ylim(c(37.5,39.5))
plot5
4.2.2 Two categorical variables (main variable and grouping variable)
Distribution of registrations per month and country
## table
tab6 <- df %>%
group_by(month, country) %>%
count() %>%
ungroup() %>%
group_by(country) %>%
mutate(perc = round(n/sum(n)*100,1))
## plot
plot6 <- ggplot(data = tab6, aes(x = month, y = perc, group = country, color = country))+
geom_point()+
geom_line(stat = "identity", size = 1, linetype = "solid")+
#geom_text(aes(label = perc), vjust = 0, hjust = 1)+
mir_theme +
scale_color_brewer(palette = "Spectral")+
labs(title = "Distribution of registrations per month and country",
x = "", y = "Percentage",
caption = "Twitter: @Shel_Kariuki")
plot6
## table
# tab6 <- df %>%
# group_by(month, country) %>%
# count() %>%
# ungroup() %>%
# group_by(country) %>%
# mutate(perc = round(n/sum(n)*100,1))
#
# ## graph
# plot6 <- ggplot(data = tab6, aes(x = month, y = perc, group = country, color = country))+
# geom_point()+
# geom_line(stat = "identity", size = 1)+
# #geom_text(aes(label = perc), hjust = 1.5, vjust = 0)+
# mir_theme+
# scale_color_brewer(palette = "Spectral")+
# labs(title = "Registrations by month",
# x = "", y = "Percentage",
# caption = "Twitter: @Shel_Kariuki")
# plot6
4.3 Scatter plot
4.3.1 2 continuous variables
Distribution of Income and household size
plot7 <- ggplot(data = df, aes(x = household_size, y = income))+
geom_point(size = 1, color = "maroon", shape = 3)+
mir_theme +
labs(title = "Distribution of Income and household size",
x = "Household Size", y = "Income",
caption = "Twitter: @Shel_Kariuki")
plot7
4.3.2 2 continuous variables and a grouping variable
plot8 <- ggplot(data = df, aes(x = household_size, y = income, color = country))+
geom_point(size = 1, shape = 3)+
mir_theme +
scale_color_manual(values = c("black","maroon","brown","yellow"))+
labs(title = "Distribution of Income and household size",
x = "Household Size", y = "Income",
caption = "Twitter: @Shel_Kariuki")
plot8
5. Other neat tricks
5.1 Reordering bar graphs by ascending order of the y axis
## table
tab2 <- df %>%
group_by(country) %>%
count() %>%
ungroup() %>%
mutate(perc = round( n/ sum(n)*100,0))
## graph
plot2b <- ggplot(data = tab2, aes(x = reorder(country,-perc), y=perc)) +
geom_bar(stat = "identity", fill = "maroon", width = 0.7) +
geom_text(aes(label = perc), size = 4 , hjust = 0.5 , vjust = -0.25)+
mir_theme +
labs(title = "Distribution of sample by country",
x = "", y = "Percentage",
caption = "Twitter: @Shel_Kariuki")+
ylim(c(0,40))
plot2b
5.2 Flipping graphs
## graph
plot2c <- ggplot(data = tab2, aes(x = country, y = perc ))+
geom_bar(stat = "identity", fill = "maroon")+
geom_text(aes(label = perc), hjust = -0.5, vjust = 0.5, size = 4.5)+
mir_theme+
labs(title = "Distribution of sample by country",
x = "", y = "Percentage",
caption = "Twitter: @Shel_Kariuki")+
coord_flip()+
ylim(c(0,40))
plot2c
5.3 Faceting plots
5.3.1 Facet wrap: faceting a plot by one variable
Distribution of education level by country
## table
tab9 <- df %>%
group_by(education_level, country) %>%
count() %>%
ungroup() %>%
group_by(country) %>%
mutate(perc = round(n/sum(n) *100,1))
##plot
plot9 <- ggplot(data = tab9, aes(x = education_level, y = perc))+
geom_bar(stat = "identity", fill = "maroon")+
geom_text(aes(label = perc), vjust = -0.25, hjust = 0.5)+
mir_theme+
theme(axis.text = element_text(size = 12))+
labs(title = "Distribution of education level by country",
x = "Education level", y = "Percentage",
caption = "Twitter: @Shel_Kariuki")+
facet_wrap(~country, ncol = 1)+
ylim(c(0,70))
plot9
5.3.2 Facet grid: Faceting a plot by two variables
Distribution of gender by country and location type
## table
tab10 <- df %>%
group_by(location_type, country, gender_of_respondent) %>%
count() %>%
ungroup() %>%
group_by(country,location_type) %>%
mutate(perc = round(n/sum(n) *100,1))
##plot
plot10 <- ggplot(data = tab10, aes(x = gender_of_respondent, y = perc))+
geom_bar(stat = "identity", fill = "maroon", width = 0.7)+
geom_text(aes(label = perc), vjust = -0.25, hjust = 0.5)+
mir_theme+
theme(axis.text = element_text(size = 12))+
labs(title = "Distribution of gender by country and location type",
x = "Gender", y = "Percentage",
caption = "Twitter: @Shel_Kariuki")+
facet_grid(country ~ location_type)+
ylim(c(0, 75))
plot10
5.4 Adding % marks on the y axis and the text labels, and altering the breaks of the y axis
## graph
plot2d <- ggplot(data = tab2, aes(x = country, y = perc ))+
geom_bar(stat = "identity", fill = "maroon")+
geom_text(aes(label = paste0(perc,"%")), hjust = 0.5, vjust = -0.25, size = 4.5)+
theme(plot.title = element_text(size = 12, #family = "Source Sans Pro Semibold",
face = "italic", hjust = 0.5),
panel.background = element_rect(fill = NA),
axis.line = element_line(size = 1, colour = "black"),
axis.text = element_text(size = 14),
axis.title = element_text(size = 16),
plot.caption = element_text(size = 14))+
ylim(c(0,40))+
scale_y_continuous(limits = c(0, 40), breaks = seq(0, 40, by = 5),
labels = function(x) paste0(x, "%"))+
labs(title = "Distribution of sample by country",
x = "", y = "Percentage",
caption = "Twitter: @Shel_Kariuki")
plot2d
5.5 Using patch work: a package for combining multiple plots
# p1 <- plot2 / plot3
# p1
end <- (plot2 + plot3) / (plot5 + plot6)
end + plot_annotation(tag_levels = "I")
