R Notebook

This is an R Markdown Notebook. When you execute code within the notebook, the results appear beneath the code.

Try executing this chunk by clicking the Run button within the chunk or by placing your cursor inside it and pressing Ctrl+Shift+Enter.

library(readxl)
library(ggplot2)
library(dplyr)
library(tidyr)

# Part C

setwd("C:/Users/kalek/Downloads")
getwd()

# Read datasets
weekends <- read.csv("C:/Users/kalek/Downloads/lisbon_weekends.csv")
weekdays <- read.csv("C:/Users/kalek/Downloads/lisbon_weekdays.csv")

cols <- c(
  "realSum", "cleanliness_rating", "guest_satisfaction_overall",
  "dist", "metro_dist",
  "attr_index", "rest_index"
)

# Summary statistics
summary_stats1 <- data.frame(
  Mean = sapply(weekends[cols], mean, na.rm = TRUE),
  SD = sapply(weekends[cols], sd, na.rm = TRUE),
  Median = sapply(weekends[cols], median, na.rm = TRUE),
  Min = sapply(weekends[cols], min, na.rm = TRUE),
  Max = sapply(weekends[cols], max, na.rm = TRUE)
)

summary_stats2 <- data.frame(
  Mean = sapply(weekdays[cols], mean, na.rm = TRUE),
  SD = sapply(weekdays[cols], sd, na.rm = TRUE),
  Median = sapply(weekdays[cols], median, na.rm = TRUE),
  Min = sapply(weekdays[cols], min, na.rm = TRUE),
  Max = sapply(weekdays[cols], max, na.rm = TRUE)
)

print("Summary Statistics for weekends:")
print(summary_stats1)

print("Summary Statistics for weekdays:")
print(summary_stats2)

columns_to_check <- c("room_type", "room_shared", "room_private", "person_capacity", 
                      "host_is_superhost", "multi", "biz", "bedrooms")

boolean_cols <- c("room_shared", "room_private", "host_is_superhost", "multi", "biz")
non_boolean_cols <- c("room_type", "person_capacity", "bedrooms")

# Function to create frequency tables
create_freq_table_df <- function(data, columns) {
  freq_tables <- list()
  
  for (col in columns) {
    freq_tbl <- as.data.frame(table(data[[col]], useNA = "ifany"))
    colnames(freq_tbl) <- c(col, "Frequency")  # Name the columns nicely
    freq_tables[[col]] <- freq_tbl
  }
  return(freq_tables)
}

for (col in non_boolean_cols) {
  unique_vals <- unique(weekends[[col]])
  cat("\n--- Column:", col, "---\n")
  cat("Number of unique values:", length(unique_vals), "\n")
  cat("Unique values:", unique_vals, "\n")
}

# For non-boolean columns
non_bool_freq_weekends <- create_freq_table_df(weekends, non_boolean_cols)
bool_freq_weekends <- create_freq_table_df(weekends, boolean_cols)

non_bool_freq_weekdays <- create_freq_table_df(weekdays, non_boolean_cols)
bool_freq_weekdays <- create_freq_table_df(weekdays, boolean_cols)

weekends$person_capacity <- as.factor(weekends$person_capacity)
weekends$bedrooms <- as.factor(weekends$bedrooms)
weekdays$person_capacity <- as.factor(weekdays$person_capacity)
weekdays$bedrooms <- as.factor(weekdays$bedrooms)

cat_vars <- c("room_type", "room_shared", "room_private", "person_capacity", 
              "multi", "biz", "bedrooms")
target_var <- "host_is_superhost"
num_vars <- c("realSum", "cleanliness_rating", "guest_satisfaction_overall",
              "dist", "metro_dist", "attr_index", "rest_index")

# 1. Chi-Square Tests: host_is_superhost vs categorical variables
cat("\n===== Chi-Square Tests (Categorical Variables) =====\n")
for (col in cat_vars) {
  tbl <- table(weekends[[col]], weekends[[target_var]])
  cat("\nChi-Square Test between", col, "and", target_var, "\n")
  print(chisq.test(tbl))
}

# 2. T-Tests: host_is_superhost vs numerical variables
cat("\n===== T-Tests (Numerical Variables) =====\n")
for (col in num_vars) {
  formula <- as.formula(paste(col, "~", target_var))
  t_test_result <- t.test(formula, data = weekends)
  cat("\nT-test for:", col, "\n")
  print(t_test_result)
}

# 3. Correlation Matrix for numerical variables
num_data <- weekends %>% select(all_of(num_vars))

cor_matrix <- cor(num_data, use = "complete.obs")

# Plot correlation matrix
ggcorrplot(cor_matrix, lab = TRUE, title = "Correlation Matrix for Numerical Attributes")

install.packages("gt")
library(ggcorrplot)
library(broom)     # for tidy test outputs
library(gt) 

chi_square_results1 <- data.frame(Weekend_Variable = character(), 
                                 Chi_Square_Statistic = numeric(), 
                                 p_value = numeric(), 
                                 stringsAsFactors = FALSE)

t_test_results1 <- data.frame(Weekend_Variable = character(), 
                             t_statistic = numeric(), 
                             p_value = numeric(), 
                             stringsAsFactors = FALSE)

for (col in cat_vars) {
  tbl <- table(weekends[[col]], weekends[[target_var]])
  chi_test <- chisq.test(tbl)
  
  chi_square_results1 <- rbind(chi_square_results1, 
                              data.frame(Variable = col,
                                         Chi_Square_Statistic = chi_test$statistic,
                                         p_value = chi_test$p.value))
}

# 2. T-Tests
for (col in num_vars) {
  formula <- as.formula(paste(col, "~", target_var))
  t_test <- t.test(formula, data = weekends)
  
  t_test_results1 <- rbind(t_test_results1, 
                          data.frame(Variable = col,
                                     t_statistic = t_test$statistic,
                                     p_value = t_test$p.value))
}


chi_square_results1 %>%
  mutate(Significant = ifelse(p_value < 0.05, "Yes", "No")) %>%
  gt() %>%
  tab_header(title = "Chi-Square Test Results (Categorical Variables)") %>%
  fmt_number(columns = c(Chi_Square_Statistic, p_value), decimals = 4)

t_test_results1 %>%
  mutate(Significant = ifelse(p_value < 0.05, "Yes", "No")) %>%
  gt() %>%
  tab_header(title = "T-Test Results (Numerical Variables)") %>%
  fmt_number(columns = c(t_statistic, p_value), decimals = 4)

num_data <- weekends %>% select(all_of(num_vars))
cor_matrix <- cor(num_data, use = "complete.obs")

ggcorrplot(cor_matrix, lab = TRUE, title = "Correlation Matrix for Numerical Attributes")

cat_vars1 <- c("room_type", "room_shared", "room_private", "person_capacity", 
              "multi", "biz", "bedrooms", "host_is_superhost")
num_vars1 <- c("cleanliness_rating", "guest_satisfaction_overall",
              "dist", "metro_dist", "attr_index", "rest_index")

anova_results1 <- data.frame(Variable = character(), 
                            F_statistic = numeric(), 
                            p_value = numeric(), 
                            stringsAsFactors = FALSE)

correlation_results1 <- data.frame(Variable = character(),
                                  Correlation = numeric(),
                                  p_value = numeric(),
                                  stringsAsFactors = FALSE)

# 1. ANOVA Tests: Categorical attributes vs realSum
for (col in cat_vars1) {
  formula <- as.formula(paste("realSum ~", col))
  anova_test <- aov(formula, data = weekends)
  anova_summary <- summary(anova_test)
  
  f_value <- anova_summary[[1]]$`F value`[1]
  p_val <- anova_summary[[1]]$`Pr(>F)`[1]
  
  anova_results1 <- rbind(anova_results1,
                         data.frame(Variable = col,
                                    F_statistic = f_value,
                                    p_value = p_val))
}

# 2. Correlation Tests: Numeric attributes vs realSum
for (col in num_vars1) {
  cor_test <- cor.test(weekends[[col]], weekends$realSum, use = "complete.obs")
  
  correlation_results1 <- rbind(correlation_results1,
                               data.frame(Variable = col,
                                          Correlation = cor_test$estimate,
                                          p_value = cor_test$p.value))
}


anova_results1 %>%
  mutate(Significant = ifelse(p_value < 0.05, "Yes", "No")) %>%
  gt() %>%
  gt::fmt_number(columns = c(F_statistic, p_value), decimals = 4) %>%
  tab_header(title = "ANOVA Test Results (Categorical Variables vs realSum)")

correlation_results1 %>%
  mutate(Significant = ifelse(p_value < 0.05, "Yes", "No")) %>%
  gt() %>%
  gt::fmt_number(columns = c(Correlation, p_value), decimals = 4) %>%
  tab_header(title = "Correlation Test Results (Numerical Variables vs realSum)")
 
num_data1 <- weekends %>% select(realSum, all_of(num_vars1))
cor_matrix1 <- cor(num_data1, use = "complete.obs")

ggcorrplot(cor_matrix1, lab = TRUE, title = "Correlation Matrix with realSum")

anova_results2 <- data.frame(Variable = character(), 
                            F_statistic = numeric(), 
                            p_value = numeric(), 
                            stringsAsFactors = FALSE)

correlation_results2 <- data.frame(Variable = character(),
                                  Correlation = numeric(),
                                  p_value = numeric(),
                                  stringsAsFactors = FALSE)

for (col in cat_vars1) {
  formula <- as.formula(paste("realSum ~", col))
  anova_test <- aov(formula, data = weekdays)
  anova_summary <- summary(anova_test)
  
  f_value <- anova_summary[[1]]$`F value`[1]
  p_val <- anova_summary[[1]]$`Pr(>F)`[1]
  
  anova_results2 <- rbind(anova_results2,
                         data.frame(Variable = col,
                                    F_statistic = f_value,
                                    p_value = p_val))
}

# 2. Correlation Tests: Numeric attributes vs realSum
for (col in num_vars1) {
  cor_test <- cor.test(weekdays[[col]], weekdays$realSum, use = "complete.obs")
  
  correlation_results2 <- rbind(correlation_results2,
                               data.frame(Variable = col,
                                          Correlation = cor_test$estimate,
                                          p_value = cor_test$p.value))
}


anova_results2 %>%
  mutate(Significant = ifelse(p_value < 0.05, "Yes", "No")) %>%
  gt() %>%
  gt::fmt_number(columns = c(F_statistic, p_value), decimals = 4) %>%
  tab_header(title = "ANOVA Test Results_weekdays (Categorical Variables vs realSum)")

correlation_results2 %>%
  mutate(Significant = ifelse(p_value < 0.05, "Yes", "No")) %>%
  gt() %>%
  gt::fmt_number(columns = c(Correlation, p_value), decimals = 4) %>%
  tab_header(title = "Correlation Test Results_weekdays (Numerical Variables vs realSum)")

num_data1i <- weekdays %>% select(realSum, all_of(num_vars1))
cor_matrix1i <- cor(num_data1i, use = "complete.obs")

ggcorrplot(cor_matrix1i, lab = TRUE, title = "Correlation Matrix with realSum")

chi_square_results2 <- data.frame(Variable = character(), 
                                 Chi_Square_Statistic = numeric(), 
                                 p_value = numeric(), 
                                 stringsAsFactors = FALSE)

t_test_results2 <- data.frame(Variable = character(), 
                             t_statistic = numeric(), 
                             p_value = numeric(), 
                             stringsAsFactors = FALSE)

for (col in cat_vars) {
  tbl <- table(weekdays[[col]], weekdays[[target_var]])
  chi_test <- chisq.test(tbl)
  
  chi_square_results2 <- rbind(chi_square_results2, 
                              data.frame(Variable = col,
                                         Chi_Square_Statistic = chi_test$statistic,
                                         p_value = chi_test$p.value))
}

# 2. T-Tests
for (col in num_vars) {
  formula <- as.formula(paste(col, "~", target_var))
  t_test <- t.test(formula, data = weekdays)
  
  t_test_results2 <- rbind(t_test_results2, 
                          data.frame(Variable = col,
                                     t_statistic = t_test$statistic,
                                     p_value = t_test$p.value))
}


chi_square_results2 %>%
  mutate(Significant = ifelse(p_value < 0.05, "Yes", "No")) %>%
  gt() %>%
  tab_header(title = "Chi-Square Test Results_weekdays (Categorical Variables)") %>%
  fmt_number(columns = c(Chi_Square_Statistic, p_value), decimals = 4)

t_test_results2 %>%
  mutate(Significant = ifelse(p_value < 0.05, "Yes", "No")) %>%
  gt() %>%
  tab_header(title = "T-Test Results_weekdays (Numerical Variables)") %>%
  fmt_number(columns = c(t_statistic, p_value), decimals = 4)


num_datai <- weekends %>% select(all_of(num_vars))
cor_matrixi <- cor(num_datai, use = "complete.obs")

ggcorrplot(cor_matrix, lab = TRUE, title = "Correlation Matrix for Numerical Attributes")

cat_vars_f <- c("room_type", "room_shared", "room_private", "person_capacity", "multi", "biz", "bedrooms", "host_is_superhost")
weekends <- weekends %>%
  mutate(across(all_of(cat_vars_f), as.factor))

weekdays <- weekdays %>%
  mutate(across(all_of(cat_vars_f), as.factor))

# 1. Build Weekends Model
weekends_model <- lm(realSum ~ room_type + room_shared + room_private + person_capacity + multi + bedrooms +
                     host_is_superhost + cleanliness_rating + guest_satisfaction_overall +
                     dist + attr_index + rest_index,
                     data = weekends)

summary(weekends_model)

weekdays_model <- lm(realSum ~ room_type + room_shared + room_private + person_capacity + multi + biz + bedrooms +
                     host_is_superhost + cleanliness_rating + guest_satisfaction_overall +
                     dist + attr_index + rest_index,
                     data = weekdays)

summary(weekdays_model)

weekdays$room_private <- as.factor(weekdays$room_private)
weekends$room_private <- as.factor(weekends$room_private)

weekends$room_type <- relevel(as.factor(weekends$room_type), ref = "Entire home/apt")

# Rebuild the model
weekends_model_final <- lm(realSum ~ room_type + room_private + person_capacity +
                           bedrooms + guest_satisfaction_overall +
                           dist + attr_index + rest_index,
                           data = weekends)

summary(weekends_model_final)

plot(predict(weekends_model_final), weekends$realSum,
     xlab = "Predicted Price", ylab = "Actual Price",
     main = "Predicted vs Actual Prices - Weekends")
abline(0, 1, col = "red")

sqrt(mean((predict(weekends_model_final) - weekends$realSum)^2))

weekends_filtered <- weekends %>%
  filter(!bedrooms %in% c(9, 10))

# Rebuild the model
weekends_model_better <- lm(realSum ~ room_type + person_capacity +
                            bedrooms + guest_satisfaction_overall +
                            dist + attr_index + rest_index,
                            data = weekends_filtered)

summary(weekends_model_better)

install.packages("modeest")
library(modeest)

columns_to_check <- c("room_type", "room_shared", "room_private", "person_capacity", 
                      "host_is_superhost", "multi", "biz", "bedrooms")

all_columns <- colnames(weekends)
numerical_variable <- setdiff(all_columns, columns_to_check)

# 1. Plot frequency graphs for categorical variables
plot_freq_graphs <- function(data, dataset_name) {
  for (var in columns_to_check) {
    ggplot(data, aes_string(x = var)) +
      geom_bar(fill = "skyblue") +
      labs(title = paste("Frequency of", var, "in", dataset_name),
           x = var, y = "Count") +
      theme_minimal() +
      theme(axis.text.x = element_text(angle = 45, hjust = 1)) -> p
    print(p)
  }
}

# 2. Plot distribution graphs for numeric variables
plot_distribution_graphs <- function(data, dataset_name) {
  for (var in numerical_variable) {
    ggplot(data, aes_string(x = var)) +
      geom_histogram(bins = 30, fill = "lightgreen", color = "black") +
      labs(title = paste("Distribution of", var, "in", dataset_name),
           x = var, y = "Frequency") +
      theme_minimal() -> p
    print(p)
  }
}

# 3. Plot min, max, mean, median, mode for numeric variables
plot_statistics_graphs <- function(data, dataset_name) {
  for (var in numerical_variable) {
    var_data <- data[[var]]
    var_data <- var_data[!is.na(var_data)]  # Remove NAs
    
    min_val <- min(var_data)
    max_val <- max(var_data)
    mean_val <- mean(var_data)
    median_val <- median(var_data)
    mode_val <- mfv(var_data)  # Most frequent value (mode)

    ggplot(data, aes_string(x = var)) +
      geom_histogram(bins = 20, fill = "lightgrey", color = "black") +
      geom_vline(xintercept = min_val, color = "blue", linetype = "dashed", size = 1, show.legend = TRUE) +
      geom_vline(xintercept = max_val, color = "red", linetype = "dashed", size = 1, show.legend = TRUE) +
      geom_vline(xintercept = mean_val, color = "green", linetype = "dashed", size = 1, show.legend = TRUE) +
      geom_vline(xintercept = median_val, color = "purple", linetype = "dashed", size = 1, show.legend = TRUE) +
      geom_vline(xintercept = mode_val, color = "orange", linetype = "dashed", size = 1, show.legend = TRUE) +
      labs(title = paste("Min/Max/Mean/Median/Mode for", var, "in", dataset_name),
           x = var, y = "Frequency") +
      theme_minimal() -> p
    print(p)
  }
}

# Run for Weekends
cat("Plotting for Weekends Dataset...\n")
plot_freq_graphs(weekends, "Weekends")
plot_distribution_graphs(weekends, "Weekends")
plot_statistics_graphs(weekends, "Weekends")

# Run for Weekdays
cat("Plotting for Weekdays Dataset...\n")
plot_freq_graphs(weekdays, "Weekdays")
plot_distribution_graphs(weekdays, "Weekdays")
plot_statistics_graphs(weekdays, "Weekdays")

weekends_private <- weekends %>% filter(room_type == "Private room")
weekends_shared <- weekends %>% filter(room_type == "Shared room")
weekends_entire <- weekends %>% filter(room_type == "Entire home/apt")

weekdays_private <- weekdays %>% filter(room_type == "Private room")
weekdays_shared <- weekdays %>% filter(room_type == "Shared room")
weekdays_entire <- weekdays %>% filter(room_type == "Entire home/apt")

# 2. Plot histogram for Private Room
ggplot(weekends_private, aes(x = realSum)) +
  geom_histogram(bins = 20, fill = "skyblue", color = "black") +
  labs(title = "Distribution of realSum - Private Room",
       x = "Price (realSum)", y = "Count") +
  theme_minimal()

# 3. Plot histogram for Shared Room
ggplot(weekends_shared, aes(x = realSum)) +
  geom_histogram(bins = 20, fill = "lightcoral", color = "black") +
  labs(title = "Distribution of realSum - Shared Room",
       x = "Price (realSum)", y = "Count") +
  theme_minimal()

# 4. Plot histogram for Entire Home/Apt
ggplot(weekends_entire, aes(x = realSum)) +
  geom_histogram(bins = 20, fill = "lightgreen", color = "black") +
  labs(title = "Distribution of realSum - Entire Home/Apt",
       x = "Price (realSum)", y = "Count") +
  theme_minimal()

# 5. Plot histogram for Private Room_weekday
ggplot(weekdays_private, aes(x = realSum)) +
  geom_histogram(bins = 20, fill = "skyblue", color = "black") +
  labs(title = "Distribution of realSum - Private Room_weekday",
       x = "Price (realSum)", y = "Count") +
  theme_minimal()

# 6. Plot histogram for Shared Room_weekday
ggplot(weekdays_shared, aes(x = realSum)) +
  geom_histogram(bins = 20, fill = "lightcoral", color = "black") +
  labs(title = "Distribution of realSum - Shared Room_weekday",
       x = "Price (realSum)", y = "Count") +
  theme_minimal()

# 7. Plot histogram for Entire Home/Apt_weekday
ggplot(weekdays_entire, aes(x = realSum)) +
  geom_histogram(bins = 20, fill = "lightgreen", color = "black") +
  labs(title = "Distribution of realSum - Entire Home/Apt_weekday",
       x = "Price (realSum)", y = "Count") +
  theme_minimal()

# Violin plot 
ggplot(weekends, aes(x = room_type, y = realSum)) +
  geom_violin(fill = "lightgreen") +
  geom_boxplot(width = 0.1, fill = "white") +
  labs(title = "Price Distribution (realSum) by Room Type (Violin Plot)",
       x = "Room Type", y = "Price (realSum)") +
  theme_minimal()

# Violin plot 
ggplot(weekdays, aes(x = room_type, y = realSum)) +
  geom_violin(fill = "lightgreen") +
  geom_boxplot(width = 0.1, fill = "white") +
  labs(title = "Price Distribution (realSum) by Room Type (Violin Plot)_weekdays",
       x = "Room Type", y = "Price (realSum)") +
  theme_minimal()

ggplot(weekends, aes(x = room_type, y = realSum)) +
  geom_boxplot(fill = "lightblue", color = "black") +
  labs(title = "Price Distribution (realSum) by Room Type",
       x = "Room Type", y = "Price (realSum)") +
  theme_minimal()

ggplot(weekdays, aes(x = room_type, y = realSum)) +
  geom_boxplot(fill = "lightblue", color = "black") +
  labs(title = "Price Distribution (realSum) by Room Type_weekdays",
       x = "Room Type", y = "Price (realSum)") +
  theme_minimal()

predict_data <- data.frame(room_type = "Private room",
                           room_shared = "False",
                           room_private = "True",
                           person_capacity = "2",
                           multi = "1",
                           bedrooms = "1",
                     host_is_superhost = "True",
                     cleanliness_rating = 10,
                     guest_satisfaction_overall = 98,
                     dist = 4.3280410,
                     attr_index = 74.23002,
                     rest_index = 176.86793)

predict_realSum <- predict(weekends_model, newdata = predict_data)

cat("Predicted price for listing with given attribute values is: $", round(predict_realSum, 2))

predict_data1 <- data.frame(room_type = "Private room",
                           room_shared = "False",
                           room_private = "True",
                           person_capacity = "3",
                           multi = "1",
                           bedrooms = "2",
                     host_is_superhost = "False",
                     cleanliness_rating = 10,
                     guest_satisfaction_overall = 100,
                     dist = 5.3260420,
                     attr_index = 64.93882,
                     rest_index = 178.96643)

predict_realSum1 <- predict(weekends_model, newdata = predict_data1)

cat("Predicted price for listing with given attribute values is: $", round(predict_realSum1, 2))

Add a new chunk by clicking the Insert Chunk button on the toolbar or by pressing Ctrl+Alt+I.

When you save the notebook, an HTML file containing the code and output will be saved alongside it (click the Preview button or press Ctrl+Shift+K to preview the HTML file).

The preview shows you a rendered HTML copy of the contents of the editor. Consequently, unlike Knit, Preview does not run any R code chunks. Instead, the output of the chunk when it was last run in the editor is displayed.