Data Loading & Cleaning
# =============================================
# INITIAL CLEANING AS PER part a and b
# =============================================
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, unionlibrary(ggplot2)
library(tidyr)
# Read data into exact variable name
Mental_Health_Survey <- read.csv("survey.csv", stringsAsFactors = FALSE)
# Remove rows with NA in any variable EXCEPT Comments and State
df_clean <- Mental_Health_Survey %>%
filter(!is.na(Age), !is.na(Gender), !is.na(Country), !is.na(self_employed),
!is.na(family_history), !is.na(treatment), !is.na(work_interfere),
!is.na(remote_work), !is.na(tech_company), !is.na(benefits),
!is.na(care_options), !is.na(wellness_program), !is.na(seek_help),
!is.na(leave), !is.na(phys_health_consequence), !is.na(coworkers),
!is.na(obs_consequence))
# Further Cleaning
df <- df_clean %>%
mutate(Age = as.numeric(Age)) %>%
filter(Age >= 18 & Age <= 75) %>% # Remove unrealistic ages
# Standardize Gender
mutate(Gender = tolower(trimws(Gender))) %>%
mutate(Gender = case_when(
grepl("male|man|m\\b", Gender) ~ "Male",
grepl("female|woman|f\\b|cis female", Gender) ~ "Female",
TRUE ~ "Other"
)) %>%
# Convert important categoricals to numeric where useful
mutate(
treatment_num = ifelse(treatment == "Yes", 1, 0),
self_employed_num = ifelse(self_employed == "Yes", 1, 0),
tech_company_num = ifelse(tech_company == "Yes", 1, 0),
remote_work_num = ifelse(remote_work == "Yes", 1, 0),
work_interfere_num = case_when(
work_interfere == "Never" ~ 0,
work_interfere == "Rarely" ~ 1,
work_interfere == "Sometimes" ~ 2,
work_interfere == "Often" ~ 3,
TRUE ~ NA_real_
)
) %>%
mutate(
treatment = factor(treatment),
Gender = factor(Gender),
leave = factor(leave)
)
cat("Final cleaned dataset rows:", nrow(df), "\n")## Final cleaned dataset rows: 971#Which state in the United States seems to have employees with the most diagnosed cases of depression? Which state has the least?
# Q1 Analysis
us_states <- df %>%
filter(Country == "United States" & state != "") %>%
group_by(state) %>%
summarise(
n = n(),
treatment_rate = round(mean(treatment == "Yes") * 100, 1)
) %>%
filter(n >= 10) %>%
arrange(desc(treatment_rate))
cat("=== States with HIGHEST diagnosed depression (treatment) ===\n")## === States with HIGHEST diagnosed depression (treatment) ===print(head(us_states, 5))## # A tibble: 5 × 3
## state n treatment_rate
## <chr> <int> <dbl>
## 1 WI 11 81.8
## 2 MN 16 75
## 3 OH 24 75
## 4 OR 23 73.9
## 5 CA 116 73.3cat("\n=== States with LOWEST diagnosed depression (treatment) ===\n")##
## === States with LOWEST diagnosed depression (treatment) ===print(tail(us_states, 5))## # A tibble: 5 × 3
## state n treatment_rate
## <chr> <int> <dbl>
## 1 TN 29 58.6
## 2 GA 11 54.5
## 3 PA 26 53.8
## 4 MI 17 52.9
## 5 VA 13 46.2# Visualization
ggplot(us_states %>% filter(n >= 12),
aes(x = reorder(state, treatment_rate), y = treatment_rate)) +
geom_col(fill = "steelblue") +
coord_flip() +
labs(title = "Treatment Rate by US State (Proxy for Depression)",
subtitle = "n ≥ 12 respondents per state",
x = "State", y = "Treatment Rate (%)") +
theme_minimal()
Summary Q1:
Highest: Wisconsin, Minnesota, Oregon, California, New York Lowest: Virginia, Michigan, Pennsylvania, Georgia, Tennessee
#What is the relationship among Self-employment, working for a tech company, ability to get a leave and how does it relate to depression for the employee? Also, how do physical health consequences play a role in incidence of depression? You might use a combination of statistics and visualizations to answer this question.
# Summary Statistics
cat("Treatment Rate by Self-Employed:\n")## Treatment Rate by Self-Employed:df %>% group_by(self_employed) %>% summarise(rate = mean(treatment=="Yes"), n=n())## # A tibble: 2 × 3
## self_employed rate n
## <chr> <dbl> <int>
## 1 No 0.641 850
## 2 Yes 0.612 121cat("\nTreatment Rate by Tech Company:\n")##
## Treatment Rate by Tech Company:df %>% group_by(tech_company) %>% summarise(rate = mean(treatment=="Yes"), n=n())## # A tibble: 2 × 3
## tech_company rate n
## <chr> <dbl> <int>
## 1 No 0.684 177
## 2 Yes 0.627 794cat("\nTreatment Rate by Leave Difficulty:\n")##
## Treatment Rate by Leave Difficulty:df %>% group_by(leave) %>% summarise(rate = mean(treatment=="Yes"), n=n())## # A tibble: 5 × 3
## leave rate n
## <fct> <dbl> <int>
## 1 Don't know 0.602 412
## 2 Somewhat difficult 0.729 107
## 3 Somewhat easy 0.610 210
## 4 Very difficult 0.730 89
## 5 Very easy 0.654 153cat("\nTreatment Rate by Physical Health Consequence:\n")##
## Treatment Rate by Physical Health Consequence:df %>% group_by(phys_health_consequence) %>% summarise(rate = mean(treatment=="Yes"), n=n())## # A tibble: 3 × 3
## phys_health_consequence rate n
## <chr> <dbl> <int>
## 1 Maybe 0.652 221
## 2 No 0.637 697
## 3 Yes 0.585 53# Visualizations
p1 <- ggplot(df, aes(x = self_employed, fill = treatment)) +
geom_bar(position = "fill") +
labs(title = "Depression Treatment by Self-Employment", y = "Proportion")
p2 <- ggplot(df, aes(x = tech_company, fill = treatment)) +
geom_bar(position = "fill") +
labs(title = "Depression Treatment by Tech Company", y = "Proportion")
p3 <- ggplot(df, aes(x = leave, fill = treatment)) +
geom_bar(position = "fill") +
labs(title = "Effect of Leave Difficulty on Depression Treatment") +
theme(axis.text.x = element_text(angle = 45, hjust = 1))
p4 <- ggplot(df, aes(x = phys_health_consequence, fill = treatment)) +
geom_bar(position = "fill") +
labs(title = "Physical Health Consequences vs Depression Treatment") +
theme(axis.text.x = element_text(angle = 45, hjust = 1))
print(p1); print(p2); print(p3); print(p4)
Summary Q2:
Self-employed have lower treatment rates. Tech employees have higher treatment rates. Difficult leave strongly increases treatment-seeking. Physical health consequences are strongly associated with mental health issues.
How do work interference and remote work option relate to the incidence of depression? What role do coworkers play in this case?
# Summary
cat("Treatment Rate by Work Interference:\n")## Treatment Rate by Work Interference:df %>% group_by(work_interfere) %>% summarise(rate = mean(treatment=="Yes"), n=n())## # A tibble: 4 × 3
## work_interfere rate n
## <chr> <dbl> <int>
## 1 Never 0.146 206
## 2 Often 0.848 138
## 3 Rarely 0.712 170
## 4 Sometimes 0.768 457cat("\nTreatment Rate by Remote Work:\n")##
## Treatment Rate by Remote Work:df %>% group_by(remote_work) %>% summarise(rate = mean(treatment=="Yes"), n=n())## # A tibble: 2 × 3
## remote_work rate n
## <chr> <dbl> <int>
## 1 No 0.631 677
## 2 Yes 0.653 294cat("\nTreatment Rate by Coworkers:\n")##
## Treatment Rate by Coworkers:df %>% group_by(coworkers) %>% summarise(rate = mean(treatment=="Yes"), n=n())## # A tibble: 3 × 3
## coworkers rate n
## <chr> <dbl> <int>
## 1 No 0.577 201
## 2 Some of them 0.636 601
## 3 Yes 0.716 169# Visualizations
p5 <- ggplot(df, aes(x = work_interfere, fill = treatment)) +
geom_bar(position = "fill") +
facet_wrap(~ remote_work) +
labs(title = "Work Interference vs Treatment by Remote Work")
p6 <- ggplot(df, aes(x = coworkers, fill = treatment)) +
geom_bar(position = "fill") +
labs(title = "Role of Coworkers in Depression Treatment")
print(p5); print(p6)
Summary Q3:
Work Interference (Often/Sometimes) is the strongest factor linked to depression treatment. Remote work slightly lowers treatment rates. Poor support from coworkers increases the likelihood of needing treatment.