R: تقرير تحليل البيانات الشامل

# 1. One line R Markdown Report
rmarkdown::render('customer_analysis.Rmd')

# ============================================
# Customer Behavior Analysis - End-to-End Projects
# Stage: 1. Data Loading -> 6. Report Output
# ============================================

library(readr)
library(dplyr)
library(tidyr)
library(stringr)
library(lubridate)
library(ggplot2)
library(broom)
library(skimr)

# ========== Stage 1: Data Loading ==========
cat('=== Stage 1: Data Loading ===\n')

set.seed(42)
n <- 2000
customers_raw <- tibble(
  customer_id = 1:n,
  age = round(rnorm(n, 35, 12)),
  gender = sample(c('M', 'F'), n, replace = TRUE),
  city = sample(c('Beijing', 'Shanghai', 'Guangzhou', 'Shenzhen', 'Hangzhou'), n, replace = TRUE),
  register_date = sample(seq(as.Date('2023-01-01'),
                              as.Date('2024-06-30'), by = 'day'), n, replace = TRUE),
  monthly_visits = round(rnorm(n, 10, 5)),
  avg_order_value = round(rnorm(n, 200, 80)),
  support_calls = sample(0:10, n, replace = TRUE),
  plan = sample(c('Basics', 'Advanced', 'Company'), n, replace = TRUE,
                 prob = c(0.5, 0.3, 0.2))
) |>
  mutate(
    logit_p = -2 + 0.02 * age - 0.05 * monthly_visits + 0.2 * support_calls,
    p = 1 / (1 + exp(-logit_p)),
    churn = rbinom(n, 1, p)
  ) |>
  select(-logit_p, -p) |>
  mutate(
    age = ifelse(row_number() %in% sample(n, 10), NA, age),
    avg_order_value = ifelse(row_number() %in% sample(n, 5), -999, avg_order_value),
    register_date = ifelse(row_number() %in% sample(n, 8), NA, register_date)
  )

write_csv(customers_raw, 'customer_raw.csv')
cat('Loaded', nrow(customers_raw), 'Raw Data Row\n\n')

# ========== Stage 2: Data Cleaning ==========
cat('=== Stage 2: Data Cleaning ===\n')

customers <- customers_raw |>
  janitor::clean_names() |>
  distinct() |>
  mutate(
    age = ifelse(is.na(age) | age < 0 | age > 150,
                 median(age[age > 0 & age < 150], na.rm = TRUE), age),
    avg_order_value = ifelse(avg_order_value < 0, NA, avg_order_value)
  ) |>
  drop_na(register_date) |>
  filter(monthly_visits >= 0, support_calls >= 0)

cat('After cleaning:', nrow(customers), 'rows\n')
cat('  - Duplicates:', nrow(customers_raw) - nrow(distinct(customers_raw)), 'removed\n')
cat('  - Age Missing/Exception: Processed\n')
cat('  - Abnormal Order Amount: Set to NA\n\n')

# ========== Stage 3: EDA Explore ==========
cat('=== Stage 3: EDA Explore ===\n')

cat('Data Dimensions:', nrow(customers), 'rows x', ncol(customers), 'cols\n')
cat('Churn rate:', round(mean(customers$churn) * 100, 2), '%\n')

churn_by_plan <- customers |>
  group_by(plan) |>
  summarise(
    n = n(),
    churn_rate = round(mean(churn) * 100, 2),
    avg_age = round(mean(age), 1),
    avg_visits = round(mean(monthly_visits), 1)
  )
cat('\nChurn Rate by Plan Type:\n')
print(churn_by_plan)

p1 <- ggplot(customers, aes(x = monthly_visits, y = avg_order_value,
                            color = factor(churn))) +
  geom_point(alpha = 0.5) +
  geom_smooth(method = 'lm', se = FALSE) +
  scale_color_brewer(palette = 'Set1', labels = c('Not lost', 'Loss')) +
  labs(title = 'Number of visits vs Average Order Value',
       x = 'Monthly Visits', y = 'Average Order Value', color = 'Status') +
  theme_minimal()

cat('Generate 2 Key Charts\n\n')

# ========== Stage 4: Descriptive Statistics ==========
cat('=== Stage 4: Descriptive Statistics ===\n')

key_metrics <- customers |>
  summarise(
    Total Number of Customers = n(),
    Number of Lost Customers = sum(churn),
    Churn rate = round(mean(churn) * 100, 2),
    Average Age = round(mean(age), 1),
    Average Monthly Visits = round(mean(monthly_visits), 1),
    Average Order Value = round(mean(avg_order_value, na.rm = TRUE), 2),
    Average Number of Customer Service Inquiries = round(mean(support_calls), 1)
  )
print(key_metrics)

by_city <- customers |>
  group_by(city) |>
  summarise(
    Number of customers = n(),
    Churn rate = round(mean(churn) * 100, 2),
    Average Order = round(mean(avg_order_value, na.rm = TRUE), 2)
  ) |>
  arrange(desc(Number of customers))
cat('\nBy City:\n')
print(by_city)
cat('\n')

# ========== Stage 5: Modeling and Analysis ==========
cat('=== Stage 5: Modeling and Analysis ===\n')

# 5.1 Logistic Regression: Churn Prediction
churn_model <- glm(churn ~ age + gender + plan + monthly_visits +
                     avg_order_value + support_calls,
                   data = customers, family = binomial)
cat('Logistic Regression Model:\n')
summary(churn_model)

cat('\nOdds Ratio Analysis:\n')
or_df <- tidy(churn_model, conf.int = TRUE, exponentiate = TRUE) |>
  filter(term != '(Intercept)')
print(or_df |> select(term, estimate, std.error, p.value, conf.low, conf.high))

# 5.3 Linear Regression: Visits Forecast
visit_model <- lm(monthly_visits ~ age + plan + avg_order_value,
                   data = customers)
cat('\nVisits Forecast Model:\n')
summary(visit_model)

# 5.4 Model Evaluation
library(pROC)
customers$churn_prob <- predict(churn_model, type = 'response')
roc_obj <- roc(customers$churn, customers$churn_prob)
cat('\nChurn Model AUC:', round(auc(roc_obj), 3), '\n\n')

# ========== Stage 6: Visualization and Reporting ==========
cat('=== Stage 6: Visualization and Reporting ===\n')

p2 <- customers |>
  group_by(plan, churn) |>
  summarise(n = n(), .groups = 'drop') |>
  ggplot(aes(x = plan, y = n, fill = factor(churn))) +
  geom_col(position = 'fill') +
  scale_y_continuous(labels = scales::percent) +
  scale_fill_brewer(palette = 'Set1', labels = c('Not lost', 'Loss')) +
  labs(title = 'Churn Rates by Plan Type', x = 'Plan', y = 'Ratio', fill = 'Status') +
  theme_minimal()

p3 <- customers |>
  group_by(plan) |>
  summarise(
    n = n(),
    churn_rate = mean(churn)
  ) |>
  ggplot(aes(x = plan, y = churn_rate, fill = plan)) +
  geom_col() +
  geom_text(aes(label = paste0(round(churn_rate * 100, 1), '%')),
            vjust = -0.5) +
  scale_y_continuous(labels = scales::percent,
                     expand = expansion(mult = c(0, 0.15))) +
  labs(title = 'Churn Rates by Plan Type', x = 'Plan', y = 'Churn rate') +
  theme_minimal() +
  theme(legend.position = 'none')

p4 <- ggplot(customers, aes(x = plan, y = monthly_visits, fill = plan)) +
  geom_boxplot() +
  labs(title = 'Distribution of Monthly Visits by Plan', x = 'Plan', y = 'Monthly Visits') +
  theme_minimal() +
  theme(legend.position = 'none')

# 6.2 Save Chart
ggsave('chart_visits_vs_order.png', p1, width = 8, height = 6, dpi = 300)
ggsave('chart_churn_by_plan.png', p2, width = 8, height = 6, dpi = 300)
ggsave('chart_churn_rate.png', p3, width = 8, height = 6, dpi = 300)
ggsave('chart_visits_by_plan.png', p4, width = 8, height = 6, dpi = 300)

# 6.3 Save Model
saveRDS(churn_model, 'churn_model.rds')
saveRDS(visit_model, 'visit_model.rds')

# 6.4 Save the processed data
write_csv(customers, 'customer_clean.csv')
saveRDS(customers, 'customer_clean.rds')

# 6.5 Summary of Business Insights
cat('\n=== Summary of Business Insights ===\n')
cat('1. Overall attrition rate:', round(mean(customers$churn) * 100, 2), '%\n')
cat('2. Customers with >5 support calls churn rate:',
    round(mean(customers$churn[customers$support_calls > 5]) * 100, 2), '%\n')
cat('3. Customers with <5 monthly visits churn rate:',
    round(mean(customers$churn[customers$monthly_visits < 5]) * 100, 2), '%\n')
cat('4. The Basic Plan has the highest churn rate (',
    round(mean(customers$churn[customers$plan == 'Basics']) * 100, 2), '%)\n')
cat('5. Older age and fewer visits -> higher probability of attrition\n')
cat('6. Suggestion: Add customers with >5 support calls to the high-risk list, proactively retain\n')

cat('\n=== All Done ===\n')
cat('Output Files:\n')
cat('  - customer_raw.csv (Raw Data)\n')
cat('  - customer_clean.csv / .rds (Data After Cleaning)\n')
cat('  - chart_*.png (4 charts)\n')
cat('  - churn_model.rds (Churn Prediction Model)\n')
cat('  - visit_model.rds (Visits Forecast Model)\n')

# Generate HTML Report
rmarkdown::render('customer_report.Rmd',
                   output_format = 'html_document',
                   output_file = 'customer_report.html')

# Generate PDF Report (LaTeX installation required)
rmarkdown::render('customer_report.Rmd',
                   output_format = 'pdf_document',
                   output_file = 'customer_report.pdf')

# ============================================
# R Tutorial 30 Lesson Study Summary Script
# After running, it outputs the topic for each lesson
# ============================================

lessons <- c(
  "01 UnderstandingR", "02 Basic Grammar", "03 Data Types", "04 Vector",
  "05 Operators", "06 Control Flow", "07 Function", "08 Matrices and Arrays",
  "09 List", "10 Data Frames and Factors",
  "11 CSV", "12 Excel",   "13 JSON and XML", "14 Database",
  "15 String", "16 Regular", "17 Data Reimagined",
  "18 Basic Drawing", "19 ggplot2 Getting Started", "20 ggplot2 Advanced",
  "21 Theme Customization", "22 Map",
  "23 Descriptive Statistics", "24 Probability Distribution", "25 Hypothesis Testing",
  "26 Linear Regression", "27 Logistic Regression",
  "28 EDA", "29 Data Cleaning", "30 Comprehensive Report"
)

cat("R Tutorial 30 Course List:\n")
for (i in seq_along(lessons)) {
  cat(sprintf("%2d. %s\n", i, lessons[i]))
}

R Tutorial 30 Course List:
 1. 01 UnderstandingR
 2. 02 Basic Grammar
...
30. 30 Comprehensive Report