Assignment 7 - Correlations - Employee Attrition Analysis

library(readr)
library(ggplot2)

hr <- read_csv('https://raw.githubusercontent.com/aiplanethub/Datasets/refs/heads/master/HR_comma_sep.csv')

## Rows: 14999 Columns: 10
## ── Column specification ────────────────────────────────────────────────────────
## Delimiter: ","
## chr (2): Department, salary
## dbl (8): satisfaction_level, last_evaluation, number_project, average_montly...
## 
## ℹ Use `spec()` to retrieve the full column specification for this data.
## ℹ Specify the column types or set `show_col_types = FALSE` to quiet this message.

Correlation 1: Satisfaction vs Last Evaluation

cor.test(hr$satisfaction_level, hr$last_evaluation) 

## 
##  Pearson's product-moment correlation
## 
## data:  hr$satisfaction_level and hr$last_evaluation
## t = 12.933, df = 14997, p-value < 2.2e-16
## alternative hypothesis: true correlation is not equal to 0
## 95 percent confidence interval:
##  0.08916727 0.12082195
## sample estimates:
##       cor 
## 0.1050212

Interpretation in technical terms

- p < 0.05 (2.2e-16), so the correlation is statistically significant.

Interpretation in non-technical terms

- If employees report higher job satisfaction, their perfomrance evaluations tend to be slightly higher too.

Scatterplot

ggplot(hr, aes(x = satisfaction_level, y = last_evaluation)) +
  geom_point(alpha = 0.15, size = 1) +
  geom_smooth(method = "lm", se = FALSE, color = "blue") +
  labs(
    title = "Higher Satisfaction is Slightly Linked to Higher Evaluations",
    x = "Satisfaction Level",
    y = "Last Evaluation Score"
  )

## `geom_smooth()` using formula = 'y ~ x'

Correlation 2: Satisfaction vs Monthly Hours

cor.test(hr$satisfaction_level, hr$average_montly_hours)

## 
##  Pearson's product-moment correlation
## 
## data:  hr$satisfaction_level and hr$average_montly_hours
## t = -2.4556, df = 14997, p-value = 0.01408
## alternative hypothesis: true correlation is not equal to 0
## 95 percent confidence interval:
##  -0.036040356 -0.004045605
## sample estimates:
##         cor 
## -0.02004811

Interpretation in technical terms

- p < 0.05 (0.01408), so the correlation is statistically significant.

Interpretation in non-technical terms

- Employees who work longer hours tend to report lower job satisfaction.

Scatterplot

ggplot(hr, aes(x = average_montly_hours, y = satisfaction_level)) +
  geom_point(alpha = 0.15, size = 1) +
  geom_smooth(method = "lm", se = FALSE, color = "blue") +
  labs(
    title = "Employees Working Longer Hours Tend to Be Less Satisfied",
    x = "Average Monthly Hours",
    y = "Satisfaction Level"
  )

## `geom_smooth()` using formula = 'y ~ x'

Correlation 3: Last Evaluation vs Average Monthly Hours

cor.test(hr$last_evaluation, hr$average_montly_hours)

## 
##  Pearson's product-moment correlation
## 
## data:  hr$last_evaluation and hr$average_montly_hours
## t = 44.237, df = 14997, p-value < 2.2e-16
## alternative hypothesis: true correlation is not equal to 0
## 95 percent confidence interval:
##  0.3255078 0.3538218
## sample estimates:
##       cor 
## 0.3397418

Interpretation in technical terms

- p < 0.05 (2.2e-16), so the correlation is statistically significant.

Interpretation in non-technical terms

- Employees who are evaluated highly usually work more hours.

Scatterplot

ggplot(hr, aes(x = average_montly_hours, y = last_evaluation)) +
  geom_point(alpha = 0.15, size  = 1) +
  geom_smooth(method = "lm", se = FALSE, color = "blue") +
  labs(
    title = "Higher Performance Scores Are Linked to More Work Hours",
    x = "Average Monthly Hours",
    y = "Last Evaluation Score"
  )

## `geom_smooth()` using formula = 'y ~ x'

Correlation 4: Time at Company vs Number of Projects

cor.test(hr$time_spend_company, hr$number_project)

## 
##  Pearson's product-moment correlation
## 
## data:  hr$time_spend_company and hr$number_project
## t = 24.579, df = 14997, p-value < 2.2e-16
## alternative hypothesis: true correlation is not equal to 0
## 95 percent confidence interval:
##  0.1813532 0.2121217
## sample estimates:
##       cor 
## 0.1967859

Interpretation in technical terms

- p < 0.05 (2.2e-16), so the correlation is statistically significant.

Interpretation in non-technical terms

- The longer someone stays at the company, the more projects they have to work on.

Scatterplot

ggplot(hr, aes(x = time_spend_company, y = number_project)) +
  geom_point(alpha = 0.4) +
  geom_smooth(method = "lm", se = FALSE, color = "blue") +
  labs(
    title = "Employees Staying Longer Tend to Work on More Projects",
    x = "Years at Company",
    y = "Number of Projects"
  )

## `geom_smooth()` using formula = 'y ~ x'