This analysis aims to investigate the relationship between demographic factors (age, gender, marital status, employment status, and monthly household income) and the likelihood of having health insurance.
To kick off, let; load the necessary libraries needed for this analysis.
library(tidyverse)
library(readxl)
library(janitor)
library(scales)
library(ggpie)
library(ggpubr)
library(ggforce)
library(htmltools)
library(plotly)
library(knitr)
library(gganimate)
library(scales)
library(kableExtra)
library(DT)
library(viridis)
library(lubridate)
library(stringr)Next, we will import the data set needed for this analysis.
health_insurance <- read_excel("D:/projects/health_insurance/health_insurance.xlsx") %>%
clean_names()After defining our problem statement, we will proceed to prepare our data for deep analysis through data cleaning procedures.
glimpse(health_insurance)Rows: 6,158
Columns: 32
$ location <chr> …
$ location_latitude <dbl> …
$ location_longitude <dbl> …
$ location_altitude <dbl> …
$ location_precision <dbl> …
$ date_and_time <dttm> …
$ age <chr> …
$ gender <chr> …
$ marital_status <chr> …
$ how_many_children_do_you_have_if_any <dbl> …
$ employment_status <chr> …
$ monthly_household_income <chr> …
$ have_you_ever_had_health_insurance <chr> …
$ if_yes_which_insurance_cover <chr> …
$ when_was_the_last_time_you_visited_a_hospital_for_medical_treatment_in_months <dbl> …
$ did_you_have_health_insurance_during_your_last_hospital_visit <chr> …
$ have_you_ever_had_a_routine_check_up_with_a_doctor_or_healthcare_provider <chr> …
$ if_you_answered_yes_to_the_previous_question_what_time_period_in_years_do_you_stay_before_having_your_routine_check_up <chr> …
$ have_you_ever_had_a_cancer_screening_e_g_mammogram_colonoscopy_etc <chr> …
$ if_you_answered_yes_to_the_previous_question_what_time_period_in_years_do_you_stay_before_having_your_cancer_screening <chr> …
$ your_picture <lgl> …
$ your_picture_url <lgl> …
$ id <dbl> …
$ uuid <chr> …
$ submission_time <dttm> …
$ validation_status <lgl> …
$ notes <lgl> …
$ status <chr> …
$ submitted_by <chr> …
$ version <chr> …
$ tags <lgl> …
$ index <dbl> …According to this, we have several columns that are not in their right data types. These columns include: age, marital status, how many children do you have and gender. To correct this, necessary data type will be allocated as follows.
health_insurance$date_and_time <- as.Date(health_insurance$date_and_time, format = "%y/%m/%d")
health_insurance$how_many_children_do_you_have_if_any <- as.numeric(health_insurance$how_many_children_do_you_have_if_any)
health_insurance$age <- as.factor(health_insurance$age)
health_insurance$marital_status <- as.factor(health_insurance$marital_status)
health_insurance$gender <- as.factor(health_insurance$gender)
#health_insurance$monthly_household_income <- as.factor(health_insurance$monthly_household_income)Next, let’s look out for missing values
missing_values <- colSums(is.na(health_insurance)) %>% print() location
353
location_latitude
353
location_longitude
353
location_altitude
353
location_precision
353
date_and_time
148
age
18
gender
17
marital_status
18
how_many_children_do_you_have_if_any
625
employment_status
24
monthly_household_income
259
have_you_ever_had_health_insurance
19
if_yes_which_insurance_cover
2519
when_was_the_last_time_you_visited_a_hospital_for_medical_treatment_in_months
158
did_you_have_health_insurance_during_your_last_hospital_visit
56
have_you_ever_had_a_routine_check_up_with_a_doctor_or_healthcare_provider
23
if_you_answered_yes_to_the_previous_question_what_time_period_in_years_do_you_stay_before_having_your_routine_check_up
4382
have_you_ever_had_a_cancer_screening_e_g_mammogram_colonoscopy_etc
31
if_you_answered_yes_to_the_previous_question_what_time_period_in_years_do_you_stay_before_having_your_cancer_screening
4593
your_picture
6158
your_picture_url
6158
id
0
uuid
0
submission_time
0
validation_status
6158
notes
6158
status
0
submitted_by
1
version
0
tags
6158
index
0 Next step is to work on the missing values.
Lets start with the “How many children do you have, if any?” column. Replace the missing values with 0 to stand for no children
health_insurance$how_many_children_do_you_have_if_any[is.na(health_insurance$how_many_children_do_you_have_if_any)] <- 0Next, fix the “Employment Status” column
health_insurance$employment_status[is.na(health_insurance$employment_status)] <- "Unknown"
health_insurance$employment_status <- as.factor(health_insurance$employment_status)The blanks in this case were replaced with “Unknown”
Next, in the “Monthly Household Income” column, we have 259 missing values. This is quite a significant number of missing values so we will replace missing values with “Unknown”
health_insurance$monthly_household_income[is.na(health_insurance$monthly_household_income)] <- "Unknown"
health_insurance$monthly_household_income <- factor(health_insurance$monthly_household_income)Next, is the “Have you ever had health insurance” column we have 19 missing values, we will replace missing values with “No”
health_insurance$have_you_ever_had_health_insurance[is.na(health_insurance$have_you_ever_had_health_insurance)] <- "No"Next up, is the “If yes, which insurance cover” column which has missing values. In this case, if the patient has insurance cover, we will replace missing value with “Unknown” but if No, we will replace missing value with “Not Applicable”
health_insurance$if_yes_which_insurance_cover <- ifelse(
health_insurance$have_you_ever_had_health_insurance == "Yes" & is.na(health_insurance$if_yes_which_insurance_cover),
"Unknown",
ifelse(
health_insurance$have_you_ever_had_health_insurance == "No",
"Not_applicable",
health_insurance$if_yes_which_insurance_cover
)
)For the “When was the last time you visited a hospital for medical treatment in months”, the missing values will be replaced with 0
health_insurance$when_was_the_last_time_you_visited_a_hospital_for_medical_treatment_in_months[is.na(health_insurance$when_was_the_last_time_you_visited_a_hospital_for_medical_treatment_in_months)] <- "0" In the “Did you have health insurance during your last hospital visit”, the missing values will be replaced with “Unknown”
health_insurance$did_you_have_health_insurance_during_your_last_hospital_visit[is.na(health_insurance$did_you_have_health_insurance_during_your_last_hospital_visit)] <- "Unknown"In the “Have you ever had a routine check up with a doctor or healthcare provider”, missing values are replaced with “Unknown”
health_insurance$have_you_ever_had_a_routine_check_up_with_a_doctor_or_healthcare_provider[is.na(health_insurance$have_you_ever_had_a_routine_check_up_with_a_doctor_or_healthcare_provider)] <- "Unknown"In the age column, we will drop the missing values. For marital status and gender, we will replace them with “Unknown”
health_insurance <- health_insurance %>%
filter(!is.na(age))health_insurance <- health_insurance %>%
mutate(gender = ifelse(is.na(gender), "Unknown", as.character(gender)))
health_insurance$gender <- factor(health_insurance$gender)health_insurance <- health_insurance %>%
mutate(marital_status = ifelse(is.na(marital_status), "Unknown", as.character(marital_status)))
health_insurance$marital_status <- factor(health_insurance$marital_status)Lastly, in the “Have you ever had a cancer screening eg.mammogram colonoscopy etc” , the missing values will be replaced with Unspecified
health_insurance$have_you_ever_had_a_cancer_screening_e_g_mammogram_colonoscopy_etc[is.na(health_insurance$have_you_ever_had_a_cancer_screening_e_g_mammogram_colonoscopy_etc)] <- "Unspecified"Next, select the necessary columns that we will need for the analysis & confirm again if there are any pending missing values
health_insurance <- health_insurance %>%
select(age,gender,marital_status, how_many_children_do_you_have_if_any, employment_status, monthly_household_income, have_you_ever_had_health_insurance,if_yes_which_insurance_cover,when_was_the_last_time_you_visited_a_hospital_for_medical_treatment_in_months, did_you_have_health_insurance_during_your_last_hospital_visit,have_you_ever_had_a_routine_check_up_with_a_doctor_or_healthcare_provider,have_you_ever_had_a_cancer_screening_e_g_mammogram_colonoscopy_etc, index)Last step in cleaning will be looking out for any duplicates in the data
duplicate <- sum(duplicated(health_insurance)) %>% print()[1] 0Now that my data is all clean, let the EDA begin!
age_group_counts <- health_insurance%>%
group_by(age) %>%
summarise(number_of_patients = n())ggplot(age_group_counts, aes(x = age, y = number_of_patients, fill = age)) +
geom_bar(stat = "identity") +
labs(
title = "Number of Patients by Age Group",
x = "Age Group",
y = "Number of Patients"
) +
theme_minimal() +
scale_fill_brewer(palette = "Set2") +
theme(legend.position = "none")
gender_group_counts <- health_insurance%>%
group_by(gender) %>%
summarise(number_of_patients = n())ggplot(gender_group_counts, aes(x = reorder(gender, -number_of_patients), y = number_of_patients)) +
geom_bar(stat = "identity", fill=c("#B45D58","#B8647D","#A875A0")) +
geom_text(aes(label = number_of_patients), hjust = 0.5, color = "black", size = 3)+
labs(title = "Gender Distribution by Number of Patients",
x = "Gender",
y = "Number of Patients") +
scale_y_continuous(labels = scales::comma_format()) +
coord_flip() +
theme_minimal() +
theme(axis.text.x = element_text(angle = 45, hjust = 1))
Findings:
We had a high record of males which was followed closely by the number of females. The margin between the two was quite small. It is also right to note that we had 10 patients that didn’t register their gender.
marital_count<- health_insurance%>%
group_by(marital_status) %>%
summarise(number_of_patients = n())ggplot(marital_count, aes(x = reorder(marital_status, -number_of_patients), y = number_of_patients)) +
geom_bar(stat = "identity", fill=c("#B45D58","#B8647D","#A875A0","#848BB8")) +
geom_text(aes(label = number_of_patients), hjust = 0.5, color = "black", size = 3)+
labs(title = "Distribution of Marital Status Among Patients",
x = "Marital Status",
y = "Number of Patients") +
scale_y_continuous(labels = scales::comma_format()) +
coord_flip() +
theme_minimal() +
theme(axis.text.x = element_text(angle = 45, hjust = 1))
Findings: The Married category had the highest number of patients whereas the Divorced category had the least number of patients. It also important to note that we have 13 patients who did not register their gender.
children_counts <- health_insurance %>%
group_by(how_many_children_do_you_have_if_any) %>%
summarise(count_of_children = n())%>%
arrange(desc(count_of_children))
datatable(children_counts,
options = list(pageLength = 10,
lengthMenu = c(10, 25, 50)),
colnames = c("Number of Children Category", "Children Count"))Findings: From the table we can conclude that the highest record was of patients with no children
employement_status<-health_insurance %>%
group_by(employment_status) %>%
summarise(employment_count=n()) %>%
arrange(desc(employment_count))ggplot(employement_status, aes(x = reorder(employment_status, -employment_count), y = employment_count)) +
geom_bar(stat = "identity", fill=c("#EF5350", "#D32F2F", "#BA68C8", "#9C27B0")) +
geom_text(aes(label = employment_count), hjust = 0.5, color = "black", size = 3)+
labs(title = "Employment Status Distribution Among Patients",
x = "Employment Status",
y = "Number of Patients") +
scale_y_continuous(labels = scales::comma_format()) +
theme_minimal() +
theme(axis.text.x = element_text(angle = 45, hjust = 1))
Findings: From the plot, it is safe to say that the highest record was of patients were employed whereas the self-employed had the least record count. However, there is a small margin between the employed and the unemployed. We also have 20 patients who didn’t record their employment status
employment_insurance2 <- health_insurance %>%
select(employment_status, have_you_ever_had_health_insurance) %>%
group_by(employment_status,have_you_ever_had_health_insurance) %>%
summarise(total=n()) %>%
ungroup() %>%
arrange(desc(total))ggplot(employment_insurance2, aes(x = reorder(employment_status, total), y = total, fill = have_you_ever_had_health_insurance)) +
geom_bar(stat = "identity", position = "dodge") +
geom_text(aes(label = total), position = position_dodge(width = 0.9), vjust = -0.5, color = "black", size = 3) +
labs(
title = "Health Insurance Status by Employment Type",
x = "Employment Status",
y = "Number of Users",
fill = "Health Insurance"
) +
scale_y_continuous(labels = scales::comma_format()) +
coord_flip() +
theme_minimal() +
theme(axis.text.x = element_text(angle = 45, hjust = 1))
employment_insurance <- health_insurance %>%
select(employment_status, have_you_ever_had_health_insurance, if_yes_which_insurance_cover) %>%
filter(have_you_ever_had_health_insurance == "Yes", employment_status %in% c("Employed","Self-employed","Unemployed")) %>%
group_by(employment_status, if_yes_which_insurance_cover) %>%
summarise(count = n(), .groups = 'drop') %>%
arrange(employment_status, desc(count)) %>%
group_by(employment_status) %>%
slice_max(order_by = count, n = 5) %>%
ungroup() %>%
mutate(employment_status = reorder(employment_status, count))ggplot(employment_insurance, aes(x = employment_status, y = count, fill = if_yes_which_insurance_cover)) +
geom_bar(stat = "identity", position = position_dodge()) +
geom_text(aes(label = count),
position = position_dodge(width = 0.9),
vjust = -0.5,
size = 3) +
labs(
title = "Top 5 Insurance Types by Employment Status",
x = "Employment Status",
y = "Count",
fill = "Insurance Type"
) +
scale_y_continuous(labels = scales::comma_format()) +
theme_minimal() +
theme(axis.text.x = element_text(angle = 45, hjust = 1))
income_count<-health_insurance %>%
group_by(monthly_household_income) %>%
summarise(income_count=n()) %>%
arrange(desc(income_count))ggplot(income_count, aes(x = reorder(monthly_household_income, -income_count), y = income_count)) +
geom_bar(stat = "identity", fill=c("#2E7D32", "#388E3C", "#4CAF50", "#81C784", "#C8E6C9",
"#FFB74D", "#FF9800")) +
geom_text(aes(label = income_count), hjust = 0.5, color = "black", size = 3)+
labs(title = "Distribution of Monthly Household Income",
x = "Income Categories",
y = "Number of Patients") +
scale_y_continuous(labels = scales::comma_format()) +
coord_flip() +
theme_minimal() +
theme(axis.text.x = element_text(angle = 45, hjust = 1))
Findings: The highest record was of patients who earn less than 10,000 where the least record count was of patients who earn 30,000-40,000. We also have 255 patients who didn’t register their monthly household income.
income_insurance2 <- health_insurance %>%
select(monthly_household_income, have_you_ever_had_health_insurance) %>%
group_by(monthly_household_income,have_you_ever_had_health_insurance) %>%
summarise(total=n()) %>%
ungroup() %>%
arrange(desc(total))ggplot(income_insurance2, aes(x = reorder(monthly_household_income, total), y = total, fill = have_you_ever_had_health_insurance)) +
geom_bar(stat = "identity", position = "dodge") +
geom_text(aes(label = total), position = position_dodge(width = 0.9), vjust = -0.5, color = "black", size = 3) +
labs(
title = "Health Insurance Status by Monthly Household Income",
x = "Monthly Household Income",
y = "Number of Users",
fill = "Health Insurance"
) +
scale_y_continuous(labels = scales::comma_format()) +
coord_flip() +
theme_minimal() +
theme(axis.text.x = element_text(angle = 45, hjust = 1))
income_insurance <- health_insurance %>%
select(monthly_household_income, have_you_ever_had_health_insurance, if_yes_which_insurance_cover) %>%
filter(have_you_ever_had_health_insurance == "Yes") %>%
group_by(monthly_household_income, if_yes_which_insurance_cover) %>%
summarise(count = n(), .groups = 'drop') %>%
arrange(monthly_household_income, desc(count)) %>%
group_by(monthly_household_income) %>%
slice_max(order_by = count, n = 3) %>%
ungroup() ggplot(income_insurance, aes(x = monthly_household_income, y = count, fill = if_yes_which_insurance_cover)) +
geom_bar(stat = "identity", position = position_dodge()) +
geom_text(aes(label = count),
position = position_dodge(width = 0.9),
vjust = -0.5,
size = 3) +
labs(
title = "Top 3 Insurance Types by Monthly Household Income",
x = "Monthly Household Income",
y = "Count",
fill = "Insurance Type"
) +
scale_y_continuous(labels = scales::comma_format()) +
theme_minimal() +
theme(axis.text.x = element_text(angle = 45, hjust = 1))
ever_had_insurance <- health_insurance %>%
group_by(have_you_ever_had_health_insurance) %>%
summarise(insurance_count=n()) %>%
arrange(desc(insurance_count))ggplot(ever_had_insurance, aes(x = reorder(have_you_ever_had_health_insurance, -insurance_count), y = insurance_count)) +
geom_bar(stat = "identity", fill=c("#EF5350", "#BA68C8")) +
geom_text(aes(label = insurance_count), hjust = 0.5, color = "black", size = 3)+
labs(title = "Have Patients Ever Had Health Insurance?",
x = "Response (Yes/No)",
y = "Number of Patients") +
scale_y_continuous(labels = scales::comma_format()) +
theme_minimal() +
theme(axis.text.x = element_text(angle = 45, hjust = 1))
#since we have insurance with the same name but different cases, we can standardize them to lower case so as to easily merge them
health_insurance <- health_insurance %>%
mutate(if_yes_which_insurance_cover = str_to_lower(if_yes_which_insurance_cover))#we also have same insurance appearing twice in different names, lets normalize it
health_insurance <- health_insurance %>%
mutate(if_yes_which_insurance_cover = case_when(
str_detect(if_yes_which_insurance_cover, regex("apa", ignore_case = TRUE)) ~ "APA Insurance",
str_detect(if_yes_which_insurance_cover, regex("jubilee", ignore_case = TRUE)) ~ "Jubilee Insurance",
str_detect(if_yes_which_insurance_cover, regex("cic", ignore_case = TRUE)) ~ "CIC Insurance",
TRUE ~ if_yes_which_insurance_cover
))insurance_type <- health_insurance %>%
group_by(if_yes_which_insurance_cover) %>%
summarise(insurance_type_count = n()) %>%
arrange(desc(insurance_type_count)) %>%
head(10)ggplot(insurance_type, aes(x = reorder(if_yes_which_insurance_cover, insurance_type_count), y = insurance_type_count, fill = insurance_type_count)) +
geom_bar(stat = "identity", fill=c("#2E7D32", "#388E3C", "#4CAF50", "#81C784", "#C8E6C9",
"#FFB74D", "#FF9800", "#F57C00", "#FF5722", "#E57373")) +
geom_text(aes(label = insurance_type_count), hjust = 0.5, color = "black", size = 3)+
labs(
title = "Top 10 Most Common Insurance Types",
x = "Insurance Type",
y = "Number of Users"
) +
scale_y_continuous(labels = scales::comma_format()) +
coord_flip() +
theme_minimal() +
theme(axis.text.x = element_text(angle = 45, hjust = 1))
Findings: According to the plot, NHIF is the most common insurance used by the patients, however, we also have a high record of patients who do not use any insurance cover (represented as “not applicable”)
hospital_cover <- health_insurance %>%
group_by(did_you_have_health_insurance_during_your_last_hospital_visit) %>%
summarise(hospital_cover_count=n()) %>%
arrange(desc(hospital_cover_count))ggplot(hospital_cover, aes(x = reorder(did_you_have_health_insurance_during_your_last_hospital_visit, -hospital_cover_count), y = hospital_cover_count)) +
geom_bar(stat = "identity", fill=c("#2E7D32", "#81C784",
"#FFB74D")) +
geom_text(aes(label = hospital_cover_count), hjust = 0.5, color = "black", size = 3)+
labs(title = "Health Insurance Coverage During Last Hospital Visit",
x = "Response (Yes or No)",
y = "Number of Patients") +
scale_y_continuous(labels = scales::comma_format()) +
coord_flip() +
theme_minimal() +
theme(axis.text.x = element_text(angle = 45, hjust = 1))
10.0 Any record of previous routine check-up with the doctor or health provider?
record_count <- health_insurance %>%
group_by(have_you_ever_had_a_routine_check_up_with_a_doctor_or_healthcare_provider) %>%
summarise(routine_check=n()) %>%
arrange(desc(routine_check))ggplot(record_count, aes(x = reorder(have_you_ever_had_a_routine_check_up_with_a_doctor_or_healthcare_provider, -routine_check), y = routine_check)) +
geom_bar(stat = "identity", fill=c("#FFB74D", "#FF9800", "#F57C00")) +
geom_text(aes(label = routine_check), hjust = 0.5, color = "black", size = 3)+
labs(title = "Records of Routine Check-Ups with Healthcare Providers",
x = "Response (Yes or No)",
y = "Number of Patients") +
scale_y_continuous(labels = scales::comma_format()) +
coord_flip() +
theme_minimal() +
theme(axis.text.x = element_text(angle = 45, hjust = 1))
Findings: Most patients do not have a history of routine check ups with healthcare providers or doctors
cancer_screening <- health_insurance %>%
group_by(have_you_ever_had_a_cancer_screening_e_g_mammogram_colonoscopy_etc) %>%
summarise(screening_count=n()) %>%
arrange(desc(screening_count))ggplot(cancer_screening, aes(x = reorder(have_you_ever_had_a_cancer_screening_e_g_mammogram_colonoscopy_etc, -screening_count), y = screening_count)) +
geom_bar(stat = "identity", fill=c( "#BA68C8", "#9C27B0", "#7B1FA2")) +
geom_text(aes(label = screening_count), hjust = 0.5, color = "black", size = 3)+
labs(title = "Cancer Screening History Among Patients",
x = "Cancer Screening (Yes/No)",
y = "Number of Patients") +
scale_y_continuous(labels = scales::comma_format()) +
coord_flip() +
theme_minimal() +
theme(axis.text.x = element_text(angle = 45, hjust = 1))
Findings: There is a very high count of patients who do not have records of cancer screenings
cancer_screen <- health_insurance %>%
select(gender, have_you_ever_had_a_cancer_screening_e_g_mammogram_colonoscopy_etc) %>%
group_by(gender, have_you_ever_had_a_cancer_screening_e_g_mammogram_colonoscopy_etc) %>%
summarise(total=n()) %>%
ungroup() %>%
print()# A tibble: 8 × 3
gender have_you_ever_had_a_cancer_screening_e_g_mammogram_colonoscopy…¹ total
<fct> <chr> <int>
1 Female No 2172
2 Female Unspecified 13
3 Female Yes 842
4 Male No 2471
5 Male Unspecified 14
6 Male Yes 618
7 Unknown No 8
8 Unknown Yes 2
# ℹ abbreviated name:
# ¹have_you_ever_had_a_cancer_screening_e_g_mammogram_colonoscopy_etcggplot(cancer_screen, aes(x = gender, y = total, fill = have_you_ever_had_a_cancer_screening_e_g_mammogram_colonoscopy_etc )) +
geom_bar(stat = "identity", position = "dodge") +
geom_text(aes(label = total),
position = position_dodge(width = 0.9),
vjust = -0.5,
color = "black",
size = 3) +
labs(
title = "Cancer Screening Status by Gender",
x = "Gender",
y = "Number of Patients",
fill = "Ever Undergone Cancer Screening?"
) +
scale_y_continuous(labels = scales::comma_format()) +
theme_minimal()
cancer_insurance <- health_insurance %>%
select(have_you_ever_had_a_cancer_screening_e_g_mammogram_colonoscopy_etc,have_you_ever_had_health_insurance, if_yes_which_insurance_cover) %>%
filter(have_you_ever_had_health_insurance == "Yes",have_you_ever_had_a_cancer_screening_e_g_mammogram_colonoscopy_etc %in% c("Yes","No")) %>%
group_by(have_you_ever_had_a_cancer_screening_e_g_mammogram_colonoscopy_etc, if_yes_which_insurance_cover) %>%
summarise(count = n(), .groups = 'drop') %>%
arrange(have_you_ever_had_a_cancer_screening_e_g_mammogram_colonoscopy_etc, desc(count)) %>%
group_by(have_you_ever_had_a_cancer_screening_e_g_mammogram_colonoscopy_etc) %>%
slice_max(order_by = count, n = 5) %>%
ungroup() ggplot(cancer_insurance, aes(x = have_you_ever_had_a_cancer_screening_e_g_mammogram_colonoscopy_etc, y = count, fill = if_yes_which_insurance_cover)) +
geom_bar(stat = "identity", position = position_dodge()) +
geom_text(aes(label = count),
position = position_dodge(width = 0.9),
vjust = -0.5,
size = 3) +
labs(
title = "Health Insurance Types by Cancer Screening Status",
x = "Cancer Screening Status (Yes/No)",
y = "Number of Patients",
fill = "Insurance Type"
) +
scale_y_continuous(labels = scales::comma_format()) +
theme_minimal() +
theme(axis.text.x = element_text(angle = 45, hjust = 1))
routine_gender <- health_insurance %>%
select(have_you_ever_had_a_routine_check_up_with_a_doctor_or_healthcare_provider, gender) %>%
group_by(gender, have_you_ever_had_a_routine_check_up_with_a_doctor_or_healthcare_provider) %>%
summarise(routine_count=n()) %>%
ungroup() %>%
arrange(desc(routine_count))ggplot(routine_gender, aes(x = gender, y = routine_count, fill = have_you_ever_had_a_routine_check_up_with_a_doctor_or_healthcare_provider )) +
geom_bar(stat = "identity", position = "dodge") +
geom_text(aes(label = routine_count),
position = position_dodge(width = 0.9),
vjust = -0.5,
color = "black",
size = 3) +
labs(
title = "Routine Check-up Status by Gender",
x = "Gender",
y = "Number of Patients",
fill = "Ever Undergone Routine Checkup?"
) +
scale_y_continuous(labels = scales::comma_format()) +
theme_minimal()
In this instance, I was looking into the type of insurance covers that people of different employment statuses went for.
routine_insurance <- health_insurance %>%
select(have_you_ever_had_a_routine_check_up_with_a_doctor_or_healthcare_provider, have_you_ever_had_health_insurance, if_yes_which_insurance_cover) %>%
filter(have_you_ever_had_health_insurance == "Yes", have_you_ever_had_a_routine_check_up_with_a_doctor_or_healthcare_provider %in% c("Yes","No")) %>%
group_by(have_you_ever_had_a_routine_check_up_with_a_doctor_or_healthcare_provider, if_yes_which_insurance_cover) %>%
summarise(count = n(), .groups = 'drop') %>%
arrange(have_you_ever_had_a_routine_check_up_with_a_doctor_or_healthcare_provider, desc(count)) %>%
group_by(have_you_ever_had_a_routine_check_up_with_a_doctor_or_healthcare_provider) %>%
slice_max(order_by = count, n = 5) %>%
ungroup() ggplot(routine_insurance, aes(x =have_you_ever_had_a_routine_check_up_with_a_doctor_or_healthcare_provider, y = count, fill = if_yes_which_insurance_cover)) +
geom_bar(stat = "identity", position = position_dodge()) +
geom_text(aes(label = count),
position = position_dodge(width = 0.9),
vjust = -0.5,
size = 3) +
labs(
title = "Health Insurance Types Among Patients by Routine Check-Up Status",
x = "Routine Check-Up Status (Yes/No)",
y = "Number of Patients",
fill = "Insurance Type"
) +
scale_y_continuous(labels = scales::comma_format()) +
theme_minimal() +
theme(axis.text.x = element_text(angle = 45, hjust = 1))