# STATLAB Academy: first analysis of the Canadian macroeconomic teaching data
# Version 1.0 - August 2026
# This script uses base R only. No package installation is required.

candidate_paths <- c(
  file.path("data", "canada_macro_monthly.csv"),
  "canada_macro_monthly.csv"
)

available_paths <- candidate_paths[file.exists(candidate_paths)]

if (length(available_paths) == 0L) {
  stop(
    paste(
      "R cannot find canada_macro_monthly.csv.",
      "Open the extracted starter-pack folder as your project,",
      "or place the CSV in the current folder.",
      "Use getwd() and list.files(recursive = TRUE) to inspect your files."
    )
  )
}

data_path <- available_paths[[1L]]
macro <- read.csv(data_path, stringsAsFactors = FALSE)

required_columns <- c(
  "date",
  "cpi_all_items_2002_100",
  "inflation_yoy_percent",
  "policy_rate_percent",
  "usd_cad_monthly_average"
)

missing_columns <- setdiff(required_columns, names(macro))
if (length(missing_columns) > 0L) {
  stop("Missing required columns: ", paste(missing_columns, collapse = ", "))
}

macro$date <- as.Date(macro$date)

if (anyNA(macro[required_columns])) {
  stop("The required data contain missing or invalid values.")
}
if (any(diff(macro$date) <= 0)) {
  stop("Dates must be unique and sorted from oldest to newest.")
}
if (nrow(macro) != 114L) {
  warning("Version 1.0 is expected to contain 114 observations.")
}

cat("\nSTATLAB Canadian macroeconomic teaching data\n")
cat("File:", data_path, "\n")
cat("Rows:", nrow(macro), "| Columns:", ncol(macro), "\n")
cat("Coverage:", format(min(macro$date)), "to", format(max(macro$date)), "\n\n")

cat("First six observations:\n")
print(head(macro))

cat("\nVariable structure:\n")
str(macro)

cat("\nSummary statistics:\n")
print(summary(macro))

latest <- tail(macro, 1L)
cat("\nLatest observation:\n")
print(latest)

numeric_names <- required_columns[-1L]
descriptive_table <- data.frame(
  variable = numeric_names,
  mean = vapply(macro[numeric_names], mean, numeric(1)),
  median = vapply(macro[numeric_names], median, numeric(1)),
  minimum = vapply(macro[numeric_names], min, numeric(1)),
  maximum = vapply(macro[numeric_names], max, numeric(1)),
  row.names = NULL
)

cat("\nSelected descriptive statistics:\n")
print(transform(
  descriptive_table,
  mean = round(mean, 3),
  median = round(median, 3),
  minimum = round(minimum, 3),
  maximum = round(maximum, 3)
))

cat("\nCorrelation matrix (association, not causation):\n")
print(round(cor(macro[numeric_names]), 3))

teaching_model <- lm(
  inflation_yoy_percent ~ policy_rate_percent + usd_cad_monthly_average,
  data = macro
)

cat("\nTeaching-only linear regression:\n")
print(summary(teaching_model))
cat(
  "\nInterpret carefully: these contemporaneous coefficients describe",
  "sample associations and are not causal effects.\n"
)

plot_file <- "canadian_inflation_plot.png"
png(plot_file, width = 1400, height = 800, res = 160)
plot(
  macro$date,
  macro$inflation_yoy_percent,
  type = "l",
  lwd = 2.5,
  col = "#08747D",
  xlab = "Month",
  ylab = "Year-over-year inflation (%)",
  main = "Canadian inflation in the STATLAB teaching dataset"
)
abline(h = 2, lty = 2, col = "#D96C4B")
legend(
  "topleft",
  legend = c("Inflation", "2% reference"),
  col = c("#08747D", "#D96C4B"),
  lty = c(1, 2),
  lwd = c(2.5, 1),
  bty = "n"
)
dev.off()

recent_file <- "canada_macro_last_24_months.csv"
write.csv(tail(macro, 24L), recent_file, row.names = FALSE)

cat("\nFiles created:\n")
cat("-", plot_file, "\n")
cat("-", recent_file, "\n")
cat("\nAnalysis complete.\n")