# Using ChickWeight Dataset

# Imagine you are monitoring the growth of young birds
  #(chicks) fed four different diets over several weeks.
  # Each chick is weighed repeatedly to see how diet
  # affects growth rate over time.

# The dataset we will use, ChickWeight,
  # mimics this kind of ecological experiment.

# Step 1: Load and Explore the Dataset
# Load the built-in dataset
ChickWeight

# To “Show me the help file (documentation) for the
  # dataset called ChickWeight.
?ChickWeight

# It contains data from an experiment on chicks fed
  # with different diets and their weight measured
  # over time.

head(ChickWeight)
str(ChickWeight)

# Step 2: Visualise Growth Trends
# We’ll plot each chick’s weight across time,
  # coloured by diet.

library(ggplot2)

ggplot(ChickWeight, aes(x = Time, y = weight, group = Chick, color = Diet)) +
  geom_line(alpha = 0.5) +
  geom_point(size = 2) +
  labs(x = "Time (days)",
       y = "Weight (g)") +
  theme_classic()

# What patterns do you observe in the growth curves
  # across diets?
# Are some diets producing faster growth than others?

# NOTE: Let's assume NORMAL DISTRIBUTION here.

# Step 3: Repeated Measures ANOVA
# We will now test:
  # (i) Whether chick weight changes over time, and
  # (ii) Whether diet affects growth patterns.

# Run repeated measures ANOVA
aov_result <- aov(weight ~ Time * Diet, data = ChickWeight)

# View the results
summary(aov_result)

# Is the effect of Time significant?
# Is the effect of Diet significant?
# Is there a significant Time × Diet interaction?
# Use the p-values to support your answer.

# Step 4: Post Hoc Test (Pairwise Comparisons)
# If the ANOVA shows significant differences
  # between diets, we can test which diets differ
  # from each other.

# For ANOVA
# If ANOVA test is significant, perform pairwise comparisons:
TukeyHSD(aov_result, "Diet")

# For Kruskal–Wallis
# If Kruskal–Wallis is significant, perform pairwise comparisons:
pairwise.wilcox.test(ChickWeight$weight, ChickWeight$Diet)

# Interpretation:
  # Compare p-values:
  # If p < 0.05, the two diets differ significantly
    # in chick weight.
  # If p > 0.05, there’s no significant difference.

# Example:
  # Diet 1 vs Diet 3 (p = 0.00048) → significant difference.
  # Diet 2 vs Diet 3 (p = 0.29) → no difference.

# Step 5: Plot Mean Growth Curves
# To visualise average growth by diet:

library(dplyr)

# Summarise mean and standard error for each diet and time
summary_data <- ChickWeight %>%
  group_by(Diet, Time) %>%
  summarise(mean_weight = mean(weight),
            se = sd(weight)/sqrt(n()), .groups = "drop")

# Plot mean ± SE
ggplot(summary_data, aes(x = Time, y = mean_weight, color = Diet, group = Diet)) +
  geom_line(size = 1.2) +
  geom_point(size = 3) +
  geom_errorbar(aes(ymin = mean_weight - se, ymax = mean_weight + se), width = 0.5) +
  labs(x = "Time (days)",
       y = "Mean Weight (g)") +
  theme_classic()

# How does this mean plot confirm or
  # contradict the earlier ANOVA results?
# Describe the growth pattern of chicks under
  # each diet.