Draft of Week 1

.github/workflows/deploy-quarto.yml

@@ -47,7 +47,7 @@ jobs:
         quarto render day1_walkthrough_answers.qmd
         quarto render day1_exercise_answers.qmd
         quarto render day2_walkthrough_answers.qmd
-#        quarto render day2_exercise_answers.qmd
+        quarto render day2_exercise_answers.qmd
     
     # Deploy HTML to GitHub Pages if on main or master branch
     - name: Deploy to GitHub pages from main/master 🚀

day2_exercise_answers.qmd

@@ -0,0 +1,276 @@
+---
+title: "Statistics in R with the `tidyverse`"
+author: "Dr. Chester Ismay"
+subtitle: "Day 2 Exercises"
+format: html
+---
+
+```{r}
+#| include: false
+options(width = 120)
+```
+
+
+# Day 2: Linear Regression in R
+
+## Session 1: Simple Linear Regression Analysis
+
+### 1. Load Necessary Libraries
+
+*We did this during the course in walkthroughs, but if you haven't done it yet:*
+
+```{r}
+# Load the required libraries
+library(dplyr)
+library(ggplot2)
+library(moderndive)
+library(readr)
+```
+
+- These packages provide tools for data wrangling, visualization, and modeling.  
+
+---
+
+### 2. Prepare Dataset
+
+```{r}
+# Load in the data_dev_survey CSV file as a data frame again
+# selecting the columns of interest keeping only complete rows of information
+data_dev_survey <- read_csv("data_dev_survey.csv") |> 
+  select(response_id, converted_comp_yearly, work_exp, 
+         years_code, ai_trust) |> 
+  # This code sets the levels of the ai_trust column to be more intuitive
+  # Note this will change the baseline level for the factor!
+  mutate(ai_trust = factor(ai_trust, levels = c("Highly distrust", 
+                                                "Somewhat distrust",
+                                                "Neither trust nor distrust",
+                                                "Somewhat trust",
+                                                "Highly trust"))) |> 
+  na.omit()
+```
+
+- We're still exploring the `data_dev_survey` data from Day 1 here.
+
+---
+
+### 3. Exploring Dataset Structure
+
+```{r}
+# Use glimpse to remind yourself of the structure of the data
+glimpse(data_dev_survey)
+```
+
+---
+
+### 4. Summary Statistics
+
+```{r}
+# Summarize mean and median values for converted_comp_yearly and work_exp
+data_dev_survey |>
+  summarize(mean_comp = mean(converted_comp_yearly),
+            median_comp = median(converted_comp_yearly),
+            mean_exp = mean(work_exp),
+            median_exp = median(work_exp))
+```
+
+---
+
+### 5. Using `tidy_summary()` Function
+
+```{r}
+# Use tidy_summary to get a clean summary of the two columns
+data_dev_survey |>
+  select(converted_comp_yearly, work_exp) |>
+  tidy_summary()
+```
+
+---
+
+### 6. Correlation Between Salary and Work Experience
+
+```{r}
+# Compute the correlation between converted_comp_yearly and work_exp
+data_dev_survey |>
+  get_correlation(formula = converted_comp_yearly ~ work_exp)
+```
+
+---
+
+### 7. Scatterplot of Salary and Work Experience
+
+```{r}
+# Plot the relationship between work experience and yearly compensation
+ggplot(data_dev_survey, aes(x = work_exp, y = converted_comp_yearly)) +
+  geom_point(alpha = 0.2) +
+  labs(x = "Work Experience (years)", y = "Yearly Compensation (USD)",
+       title = "Scatterplot of Work Experience and Yearly Compensation")
+```
+
+---
+
+### 8. Regression Line
+
+```{r}
+# The same scatterplot with a regression line added
+ggplot(data_dev_survey, aes(x = work_exp, y = converted_comp_yearly)) +
+  geom_point(alpha = 0.2) +
+  geom_smooth(method = "lm", se = FALSE) +
+  labs(x = "Work Experience (years)", y = "Yearly Compensation (USD)",
+       title = "Work Experience vs. Yearly Compensation with Regression Line")
+```
+
+---
+
+### 9. Fit a Regression Model
+
+```{r}
+# Fit a linear regression model with compensation as the outcome
+salary_slr_model <- lm(converted_comp_yearly ~ work_exp, data = data_dev_survey)
+
+# Get the regression coefficients
+coef(salary_slr_model)
+```
+
+---
+
+## Session 2: Multiple Linear Regression Analysis (Part 1)
+
+### 10: Fitting a Simple Linear Regression with a Categorical Regressor
+
+```{r}
+# Fit a simple linear regression model with ai_trust as a predictor on salary
+trust_model <- lm(converted_comp_yearly ~ ai_trust, data = data_dev_survey)
+
+# Get regression coefficients
+coef(trust_model)
+```
+
+---
+
+### 11: Visualizing Data with Categorical Variable
+
+```{r}
+# Scatterplot with ai_trust as color for grouping on work_exp predicting
+# converted_comp_yearly
+ggplot(data_dev_survey, 
+       aes(x = work_exp, y = converted_comp_yearly, color = ai_trust)) +
+  geom_point(alpha = 0.5) +
+  labs(x = "Work Experience (years)", y = "Yearly Compensation (USD)",
+       color = "Trust in AI",
+       title = "Scatterplot of Work Experience and Yearly Compensation by Trust in AI")
+```
+
+---
+
+### 12: Summarizing the Target and Regressors (One numerical and one categorical)
+
+```{r}
+# Summarize the target and regressors
+data_dev_survey |>
+  select(converted_comp_yearly, work_exp, ai_trust) |>
+  tidy_summary()
+```
+
+---
+
+### 13: Fitting a Multiple Regression Model with Interaction Terms
+
+```{r}
+# Fit a multiple regression model with interaction
+multi_model_interaction <- lm(converted_comp_yearly ~ work_exp * ai_trust, 
+                              data = data_dev_survey)
+
+# Get regression coefficients
+coef(multi_model_interaction)
+
+```
+
+---
+
+### 14: Adding Interaction Terms to the Regression Plot
+
+```{r}
+# Scatterplot with regression lines by trust in AI
+ggplot(data_dev_survey, aes(x = work_exp, y = converted_comp_yearly, color = ai_trust)) +
+  geom_point(alpha = 0.5) +
+  geom_smooth(method = "lm", se = FALSE) +
+  labs(x = "Work Experience (years)", y = "Yearly Compensation (USD)",
+       color = "Trust in AI",
+       title = "Work Experience vs. Yearly Compensation by Trust in AI")
+```
+
+---
+
+## Session 3: Multiple Linear Regression Analysis (Part 2)
+
+### 15. Fitting a Multiple Regression Model Without Interactions
+
+```{r}
+# Fit a multiple regression model without interaction terms
+multi_model_no_interaction <- lm(converted_comp_yearly ~ work_exp + ai_trust, 
+                                 data = data_dev_survey)
+
+# Get regression coefficients
+coef(multi_model_no_interaction)
+```
+
+---
+
+### 16. Visualizing the Parallel Slopes Model
+
+```{r}
+# Scatterplot with regression lines by trust in AI (parallel slopes)
+ggplot(data_dev_survey, aes(x = work_exp, y = converted_comp_yearly, color = ai_trust)) +
+  geom_point(alpha = 0.5) +
+  geom_parallel_slopes(se = FALSE) +
+  labs(x = "Work Experience (years)", y = "Yearly Compensation (USD)",
+       color = "Trust in AI",
+       title = "Work Experience vs. Yearly Compensation by Trust in AI")
+```
+
+---
+
+### 17. Summarize the Numerical Regressors and Target Variable
+
+```{r}
+# Use tidy_summary to get a clean summary of the data for a multiple regression
+# with work experience and years coding as regressors and yearly compensation as
+# the outcome
+data_dev_survey |>
+  select(converted_comp_yearly, work_exp, years_code) |>
+  tidy_summary()
+```
+
+---
+
+### 18. Fitting a Multiple Regression Model with Two Numerical Predictors
+
+```{r}
+# Fit a multiple regression model
+multi_model <- lm(converted_comp_yearly ~ work_exp + years_code, 
+                  data = data_dev_survey)
+
+# Get regression coefficients
+coef(multi_model)
+```
+
+---
+
+### 19. Visualizing the Relationship between Variables (with SLR)
+
+```{r}
+# Create scatterplots with regression lines for both variables with 
+# converted_comp_yearly as the response variable in simple linear regression
+ggplot(data_dev_survey, aes(x = work_exp, y = converted_comp_yearly)) +
+  geom_point(alpha = 0.5) +
+  geom_smooth(method = "lm", se = FALSE) +
+  labs(x = "Work Experience (years)", y = "Yearly Compensation (USD)",
+       title = "Yearly Compensation vs Work Experience")
+
+ggplot(data_dev_survey, aes(x = years_code, y = converted_comp_yearly)) +
+  geom_point(alpha = 0.5) +
+  geom_smooth(method = "lm", se = FALSE) +
+  labs(x = "Years Coding", y = "Yearly Compensation (USD)",
+       title = "Yearly Compensation vs Years Coding")
+```
+