Pakistan’s Exports: What Sells and Where It Goes

A self-updating report, pulled live from the SBP EasyData API at every render

NoteThe story so far (data to August 2026)

In August 2026, Pakistan exported goods worth $2.46 billion, down 1.1% on August 2025. Over the last twelve months goods exports reached $31.03 billion, down 5.3% on the year before. FY2025-26 closed at $30.82 billion (down 4.7%). The first 2 months of FY2026-27 brought in $5.44 billion, up 4.0% on the same months of FY2025-26. Services added another $10.43 billion over the last twelve months, 25.2% of total exports.

Every number above is calculated at render time from data downloaded from SBP EasyData on 22 September 2026. Hover over any chart for exact values; drag the slider under a chart to zoom into a period.

About the data

Goods and services exports come from SBP’s monthly balance of payments summary (BPM6), in million US dollars.

Goods exports

Figure 1

Figure 2

The bars swing with Eid, working days and shipment timing, so the rolling line is the better guide to direction.

Figure 3

Figure 4: Cumulative goods exports through the fiscal year

This chart puts the current fiscal year beside the three before it, month by month, so it shows at a glance whether the year is running ahead or behind.

Services exports

Figure 5

Figure 6

Figure 7

Fiscal year summary

Table 1: Growth is shown only between complete fiscal years

Fiscal year	Goods (US$ m)	Goods growth	Services (US$ m)	Services growth	Services share of total
FY2013-14	25078	NA	5345	NA	17.6%
FY2014-15	24090	-3.9%	5872	9.9%	19.6%
FY2015-16	21972	-8.8%	5456	-7.1%	19.9%
FY2016-17	22003	0.1%	5915	8.4%	21.2%
FY2017-18	24768	12.6%	5851	-1.1%	19.1%
FY2018-19	24257	-2.1%	5966	2.0%	19.7%
FY2019-20	22536	-7.1%	5437	-8.9%	19.4%
FY2020-21	25639	13.8%	5945	9.3%	18.8%
FY2021-22	32493	26.7%	7102	19.5%	17.9%
FY2022-23	27876	-14.2%	7596	7.0%	21.4%
FY2023-24	30980	11.1%	7691	1.3%	19.9%
FY2024-25	32343	4.4%	8450	9.9%	20.7%
FY2025-26	30821	-4.7%	10020	18.6%	24.5%
FY2026-27 (to date)	5445	NA	1811	NA	25.0%

The last twelve months

Table 2

Month	Goods (US$ m)	Goods, vs a year earlier	Services (US$ m)	Services, vs a year earlier
Sep 2025	2609	-0.3%	807	23.8%
Oct 2025	2632	-12.4%	811	15.2%
Nov 2025	2277	-18.4%	797	11.6%
Dec 2025	2758	-11.3%	946	17.2%
Jan 2026	2749	-8.6%	868	28.8%
Feb 2026	2481	-4.9%	799	14.8%
Mar 2026	2528	-8.4%	903	16.2%
Apr 2026	2619	0.5%	904	20.4%
May 2026	2357	-3.6%	838	15.7%
Jun 2026	2573	-0.7%	942	35.0%
Jul 2026	2985	8.5%	939	29.0%
Aug 2026	2460	-1.1%	872	28.8%

Keeping it current

Each render downloads every series again, so quarto render exports_sbp.qmd after SBP’s monthly release updates every number, chart and sentence. SBP occasionally fails a request, so each series is tried up to four times. Destination and product sections appear automatically once their series codes are listed in sbp_export_series.csv (build it with -P rediscover:true); until then the report covers goods and services only.

Table 3

Appendix: code

library(tidyverse)
library(httr2)
library(tsibble)
library(knitr)
library(plotly)
library(plotly)

if (!exists("params")) params <- list(rediscover = FALSE)
theme_set(theme_minimal(base_size = 12))

start_date <- "2001-07-01"
keys_file  <- "sbp_export_series.csv"   # series codes and names only, no data

# Headline exports, BPM6 monthly summary
headline_keys <- c(
  goods    = "TS_GP_BOP_BPM6SUM_M.P00030",   # Exports of goods FOB
  services = "TS_GP_BOP_BPM6SUM_M.P00060"    # Exports of services
)

# Breakdowns and candidate series-code prefixes (used only when scanning)
blocks <- tribble(
  ~block,       ~dataset,               ~prefixes,
  "country",    "TS_GP_BOP_XRECCOU_M",  list(c("EXPC", "EXPA")),   # EXPC mirrors the confirmed IMPC
  "product",    "TS_GP_BOP_XRECCOM_M",  list(c("EXPA", "EXPC"))    # EXPA confirmed: ...XRECCOM_M.EXPA00420
)

block_title <- c(country = "export destinations", product = "export products")

country_agg <- regex(str_c(
  "\\btotal\\b|countries|region|union|world|unspecified|not specified|n\\.e\\.s|",
  "^(asia|africa|europe|oceania|america)\\b|^(north|south|latin|central) america|",
  "^(east|west|south|central|north)(ern)? (asia|europe)|middle east|far east|",
  "\\bgcc\\b|saarc|asean|opec|commonwealth|^others?\\b"
), ignore_case = TRUE)

product_agg <- regex("\\btotal\\b|^others?\\b|\\bgroup\\b", ignore_case = TRUE)
sbp_key <- Sys.getenv("SBP_EASYDATA_KEY")
if (!nzchar(sbp_key)) {
  stop("SBP_EASYDATA_KEY not found. Add it with usethis::edit_r_environ() and restart R.")
}

sbp_req <- function(key, start = start_date, end = Sys.Date(), timeout = 30) {
  request("https://easydata.sbp.org.pk/api/v1/series") |>
    req_url_path_append(key, "data") |>
    req_url_query(api_key = sbp_key, format = "json",
                  start_date = as.character(as.Date(start)),
                  end_date   = as.character(as.Date(end))) |>
    req_headers(Accept = "application/json") |>
    req_user_agent("pakistan-export-report (R httr2)") |>
    req_timeout(timeout) |>
    req_error(is_error = \(r) FALSE)
}

# SBP JSON: { columns: [...], rows: [[dataset, key, name, date, value, unit, ...], ...] }
parse_sbp <- function(resp, key) {
  body <- resp_body_json(resp, simplifyVector = FALSE)
  if (is.null(body$rows) || length(body$rows) == 0) return(NULL)

  cols <- str_to_lower(unlist(body$columns))
  find <- function(pattern, fallback) {
    i <- which(str_detect(cols, pattern))
    if (length(i) == 0) fallback else i[1]
  }
  i_name <- find("series name|^series$", 3)
  i_date <- find("observation date", 4)
  i_val  <- find("observation value", 5)
  cell   <- function(r, i) if (length(r) >= i && !is.null(r[[i]])) as.character(r[[i]]) else NA_character_

  tibble(
    key         = key,
    series_name = map_chr(body$rows, cell, i = i_name),
    date        = map_chr(body$rows, cell, i = i_date),
    value       = map_chr(body$rows, cell, i = i_val)
  ) |>
    mutate(
      month = yearmonth(as.Date(parse_date_time(date, c("Ymd", "dbY", "dby", "dmY"), quiet = TRUE))),
      value = parse_number(value)
    ) |>
    filter(!is.na(month), !is.na(value)) |>
    select(key, series_name, month, value)
}
safe_parse <- possibly(parse_sbp, otherwise = NULL)

empty_pull <- tibble(key = character(), series_name = character(),
                     month = yearmonth(as.Date(character())), value = numeric())

# One series at a time, with patient retries. SBP sometimes fails a call and
# answers the same call a few seconds later, so each series gets up to
# `attempts` tries with growing waits.
pull_keys <- function(keys, start = start_date, timeout = 30, attempts = 4,
                      waits = c(2, 5, 10), pause = 0.8,
                      deadline = Inf, verbose = TRUE) {
  keys <- unique(keys)
  if (length(keys) == 0) return(empty_pull)
  out <- vector("list", length(keys))

  for (i in seq_along(keys)) {
    note <- ""
    for (a in seq_len(attempts)) {
      if (Sys.time() > deadline) break
      Sys.sleep(if (a == 1) pause else waits[min(a - 1, length(waits))])
      resp <- tryCatch(req_perform(sbp_req(keys[i], start = start, timeout = timeout)),
                       error = \(e) e)

      if (!inherits(resp, "httr2_response")) {
        note <- str_c("request failed: ", conditionMessage(resp))
      } else if (resp_status(resp) != 200) {
        note <- str_c("HTTP ", resp_status(resp))
        if (resp_status(resp) == 404) break          # wrong code, no point retrying
      } else {
        out[[i]] <- safe_parse(resp, keys[i])
        if (!is.null(out[[i]])) { note <- str_c("ok, ", nrow(out[[i]]), " rows"); break }
        note <- "HTTP 200 but no usable rows"
      }
    }
    if (verbose) message(sprintf("  [%d/%d] %s  %s (attempt %d)",
                                 i, length(keys), keys[i], note, a))
    if (Sys.time() > deadline) { message("  time limit reached, stopping"); break }
  }
  out <- list_rbind(out)
  if (is.null(out) || ncol(out) == 0) empty_pull else out
}
# Scanning is OFF by default. Run it once with -P rediscover:true, or write
# sbp_export_series.csv by hand with columns: block, key, series_name.

code_keys <- function(dataset, prefix, width, codes) {
  str_c(dataset, ".", prefix, str_pad(codes, width, pad = "0"))
}

# Walk the codes upward in steps of 10. Stop when there has been no hit for a
# long stretch, so the scan ends by itself once the list of series runs out.
scan_prefix <- function(dataset, prefix, deadline, width = 5, step = 10,
                        give_up_before_first = 80, give_up_after_last = 40) {
  recent <- Sys.Date() - 540
  hits <- list(); misses <- 0; code <- step; any_hit <- FALSE
  repeat {
    if (Sys.time() > deadline) { message("  time limit reached"); break }
    key <- code_keys(dataset, prefix, width, code)
    got <- pull_keys(key, start = recent, timeout = 10, attempts = 2,
                     waits = 2, pause = 0.3, verbose = FALSE)
    if (nrow(got) > 0) {
      any_hit <- TRUE; misses <- 0
      hits[[length(hits) + 1]] <- got |> summarise(series_name = last(series_name), .by = key)
      message("  ", key, "  ", last(got$series_name))
    } else {
      misses <- misses + 1
    }
    if (!any_hit && misses >= give_up_before_first) break
    if (any_hit && misses >= give_up_after_last) break
    code <- code + step
  }
  list_rbind(hits)
}

discover_block <- function(dataset, prefixes, deadline) {
  for (p in prefixes) {
    message("Scanning ", dataset, ".", p, "...")
    found <- scan_prefix(dataset, p, deadline)
    if (!is.null(found) && nrow(found) > 0) return(found)
  }
  NULL
}

if (isTRUE(params$rediscover)) {
  deadline <- Sys.time() + 20 * 60
  message("Scanning SBP for export series (max 20 minutes)...")
  found <- blocks |>
    mutate(found = map2(dataset, prefixes, \(d, p) discover_block(d, p, deadline))) |>
    select(block, found) |>
    unnest(found)
  if (nrow(found) > 0) write_csv(found, keys_file)
}

directory <- if (file.exists(keys_file)) {
  read_csv(keys_file, show_col_types = FALSE)
} else {
  tibble(block = character(), key = character(), series_name = character())
}
raw <- pull_keys(c(headline_keys, directory$key))
pulled_on <- format(Sys.Date(), "%d %B %Y")

if (!headline_keys[["goods"]] %in% raw$key) {
  stop("SBP did not return goods exports (", headline_keys[["goods"]], ") after several tries. ",
       "The messages above show what SBP answered; please render again in a few minutes.")
}

roll12 <- function(x) as.numeric(stats::filter(x, rep(1, 12), sides = 1))

# ---- Headline exports ----
exp_m <- raw |>
  filter(key %in% headline_keys) |>
  mutate(series = names(headline_keys)[match(key, headline_keys)], value = abs(value)) |>
  distinct(month, series, .keep_all = TRUE) |>
  select(month, series, value) |>
  pivot_wider(names_from = series, values_from = value) |>
  (\(d) { if (!"services" %in% names(d)) d$services <- NA_real_; d })() |>
  arrange(month) |>
  filter(!is.na(goods)) |>
  mutate(
    goods12  = roll12(goods),
    serv12   = roll12(services),
    goods_yoy   = goods / lag(goods, 12) - 1,
    goods12_yoy = goods12 / lag(goods12, 12) - 1,
    serv_yoy    = services / lag(services, 12) - 1,
    serv12_yoy  = serv12 / lag(serv12, 12) - 1,
    serv12   = roll12(services),
    fy       = year(as.Date(month)) + (month(as.Date(month)) >= 7),
    fy_month = (month(as.Date(month)) - 7) %% 12 + 1
  )

# ---- Destinations and products ----
detail <- raw |>
  inner_join(select(directory, key, block), by = join_by(key)) |>
  mutate(label = str_squish(series_name)) |>
  distinct(key, month, .keep_all = TRUE)

if (nrow(detail) > 0) {
  series_info <- detail |>
    summarise(mean_v = mean(value), .by = c(block, key, label)) |>
    mutate(
      named_total = str_detect(label, regex("\\btotal\\b", ignore_case = TRUE)),
      others_sum  = sum(mean_v) - mean_v,
      .by = block
    ) |>
    mutate(
      total_key = {
        cand <- key[named_total]
        if (length(cand) > 0) cand[which.max(mean_v[named_total])]
        else if (max(mean_v) > 0.5 * others_sum[which.max(mean_v)]) key[which.max(mean_v)]
        else NA_character_
      },
      total_mean = if (all(is.na(total_key))) NA_real_ else mean_v[key == first(total_key)],
      .by = block
    ) |>
    mutate(role = case_when(
      key == total_key ~ "Total",
      block == "country" & str_detect(label, country_agg) ~ "Aggregate",
      block == "product" & str_detect(label, product_agg) ~ "Aggregate",
      !is.na(total_mean) & mean_v > 0.5 * total_mean ~ "Aggregate",
      .default = "Item"
    ))

  items <- detail |>
    semi_join(filter(series_info, role == "Item"), by = join_by(key)) |>
    summarise(value = sum(value), .by = c(block, label, month))

  block_totals <- bind_rows(
    detail |>
      semi_join(filter(series_info, role == "Total"), by = join_by(key)) |>
      select(block, month, value),
    items |>
      filter(!block %in% series_info$block[series_info$role == "Total"]) |>
      summarise(value = sum(value), .by = c(block, month))
  )
} else {
  series_info  <- tibble(block = character(), key = character(), label = character(),
                         mean_v = numeric(), role = character())
  items        <- tibble(block = character(), label = character(),
                         month = yearmonth(as.Date(character())), value = numeric())
  block_totals <- tibble(block = character(), month = yearmonth(as.Date(character())), value = numeric())
}

has <- as.list(set_names(names(block_title) %in% unique(items$block), names(block_title)))
usd <- function(x) {
  x <- unname(x)
  if_else(abs(x) >= 1000,
          str_c("$", format(round(abs(x) / 1000, 2), nsmall = 2), " billion"),
          str_c("$", format(round(abs(x)), big.mark = ","), " million"))
}
pct   <- function(x, d = 1) str_c(format(round(100 * abs(unname(x)), d), nsmall = d), "%")
chg   <- function(x) if_else(is.na(x), "new", str_c(if_else(x >= 0, "up ", "down "), pct(x)))
mlab  <- function(m) format(as.Date(m), "%B %Y")
fylab <- function(fy) str_c("FY", fy - 1, "-", sprintf("%02d", fy %% 100))
list_and <- function(x) {
  if (length(x) <= 1) return(x)
  str_c(str_c(head(x, -1), collapse = ", "), " and ", tail(x, 1))
}

analyse_block <- function(blk, n = 10) {
  tot <- block_totals |> filter(block == blk)
  if (nrow(tot) == 0) return(NULL)
  latest_b <- max(tot$month)
  in_now   <- function(m) m > latest_b - 12 & m <= latest_b
  in_prev  <- function(m) m > latest_b - 24 & m <= latest_b - 12
  tot_now  <- sum(tot$value[in_now(tot$month)])
  tot_prev <- sum(tot$value[in_prev(tot$month)])

  tbl <- items |>
    filter(block == blk) |>
    summarise(now = sum(value[in_now(month)]), before = sum(value[in_prev(month)]), .by = label) |>
    mutate(
      growth       = if_else(before > 0, now / before - 1, NA_real_),
      share        = now / tot_now,
      share_before = before / tot_prev,
      rank_now     = min_rank(desc(now)),
      rank_before  = min_rank(desc(before))
    ) |>
    arrange(rank_now, label)

  top <- slice_head(tbl, n = n)
  list(
    block = blk, latest = latest_b, tot_now = tot_now,
    table = tbl, top = top,
    top_share = sum(top$share),
    top_share_before = sum(slice_min(tbl, rank_before, n = n, with_ties = FALSE)$share_before),
    newcomers = top$label[top$rank_before > n]
  )
}

A <- map(set_names(names(block_title)), \(b) if (isTRUE(has[[b]])) analyse_block(b) else NULL)

block_msg <- function(a) {
  if (is.null(a)) return("")
  what   <- block_title[[a$block]]
  ranked <- filter(a$top, !is.na(growth))
  fast   <- slice_max(ranked, growth, n = 1, with_ties = FALSE)
  slow   <- slice_min(ranked, growth, n = 1, with_ties = FALSE)
  str_c(
    str_glue(
      "In the twelve months to {mlab(a$latest)}, the top 10 {what} accounted for {pct(a$top_share)} ",
      "of the {usd(a$tot_now)} total ({if_else(a$top_share >= a$top_share_before, 'up', 'down')} from ",
      "{pct(a$top_share_before)} a year earlier). {a$top$label[1]} leads with {usd(a$top$now[1])}, ",
      "a {pct(a$top$share[1])} share, followed by {a$top$label[2]} ({pct(a$top$share[2])}) and ",
      "{a$top$label[3]} ({pct(a$top$share[3])}). Within the top 10, {fast$label} grew fastest ",
      "({chg(fast$growth)}) while {slow$label} ",
      "{if_else(slow$growth < 0, 'fell furthest', 'grew least')} ({chg(slow$growth)}). "
    ),
    if (length(a$newcomers) > 0) {
      str_glue("{list_and(a$newcomers)} {if (length(a$newcomers) > 1) 'are' else 'is'} new to the top 10 this year.")
    } else {
      "The top 10 line-up is unchanged from the previous year."
    }
  )
}

plot_top <- function(a) {
  if (is.null(a)) return(invisible(NULL))
  a$top |>
    mutate(label = fct_reorder(label, now),
           tag = str_c(pct(share), "  ·  ", chg(growth))) |>
    ggplot(aes(now / 1000, label,
               text = str_c(label, "<br>", usd(now), "<br>Share: ", pct(share),
                            "<br>Change: ", chg(growth)))) +
    geom_col(fill = "#1b5e20", width = 0.7) +
    geom_text(aes(label = tag), hjust = -0.05, size = 3.2, colour = "grey30") +
    scale_x_continuous(expand = expansion(mult = c(0, 0.35))) +
    labs(x = str_c("US$ billion, 12 months to ", mlab(a$latest)), y = NULL,
         caption = "Labels: share of total · change on the previous 12 months")
}

plot_trend <- function(a) {
  if (is.null(a)) return(invisible(NULL))
  items |>
    filter(block == a$block, label %in% a$top$label) |>
    select(label, month, value) |>
    complete(label, month = seq(min(month), max(month), by = 1), fill = list(value = 0)) |>
    arrange(label, month) |>
    mutate(r12 = roll12(value), .by = label) |>
    filter(!is.na(r12)) |>
    mutate(label = factor(label, levels = a$top$label)) |>
    ggplot(aes(as.Date(month), r12 / 1000)) +
    geom_line(colour = "#1b5e20", linewidth = 0.8) +
    facet_wrap(~label, ncol = 5, scales = "free_y") +
    labs(x = NULL, y = "US$ billion, rolling 12 months") +
    theme(strip.text = element_text(face = "bold", size = 9),
          axis.text.x = element_text(size = 7))
}

table_top <- function(a) {
  if (is.null(a)) return(invisible(NULL))
  a$top |>
    transmute(
      Rank = rank_now,
      Name = label,
      `12 months (US$ m)` = round(now),
      `Previous 12 months (US$ m)` = round(before),
      Change = if_else(is.na(growth), "new", scales::percent(growth, accuracy = 0.1)),
      Share = scales::percent(share, accuracy = 0.1),
      `Rank a year ago` = rank_before
    ) |>
    kable()
}
latest  <- max(exp_m$month)
lr      <- exp_m |> filter(month == latest)
ly      <- exp_m |> filter(month == latest - 12)
cur_fy  <- lr$fy
cur_fm  <- lr$fy_month
fytd    <- sum(exp_m$goods[exp_m$fy == cur_fy])
fytd_ly <- sum(exp_m$goods[exp_m$fy == cur_fy - 1 & exp_m$fy_month <= cur_fm])

fy_done <- exp_m |>
  summarise(goods = sum(goods), months = n(), .by = fy) |>
  filter(months == 12) |>
  arrange(fy) |>
  mutate(growth = goods / lag(goods) - 1)
last_fy <- slice_tail(fy_done, n = 1)

svc_ok  <- !all(is.na(exp_m$services))
svc12   <- if (svc_ok) sum(tail(exp_m$services, 12), na.rm = TRUE) else NA

msg_head <- str_c(
  str_glue(
    "In {mlab(latest)}, Pakistan exported goods worth {usd(lr$goods)}, ",
    "{chg(lr$goods / ly$goods - 1)} on {mlab(latest - 12)}. Over the last twelve months goods ",
    "exports reached {usd(lr$goods12)}, {chg(lr$goods12 / ly$goods12 - 1)} on the year before. ",
    "{fylab(last_fy$fy)} closed at {usd(last_fy$goods)} ({chg(last_fy$growth)})"
  ),
  if (last_fy$goods == max(fy_done$goods)) ", the highest annual total in this series." else ".",
  if (cur_fm < 12) {
    str_glue(" The first {cur_fm} month{if (cur_fm > 1) 's' else ''} of {fylab(cur_fy)} ",
             "brought in {usd(fytd)}, {chg(fytd / fytd_ly - 1)} on the same months of {fylab(cur_fy - 1)}.")
  } else "",
  if (svc_ok) {
    str_glue(" Services added another {usd(svc12)} over the last twelve months, ",
             "{pct(svc12 / (svc12 + lr$goods12))} of total exports.")
  } else ""
)
hover_m <- "%{x|%b %Y}<br>US$ %{y:,.0f} million<extra>%{fullData.name}</extra>"

monthly_chart <- function(df, value, avg, bar_col, line_col) {
  plot_ly(df, x = ~as.Date(month)) |>
    add_bars(y = value, name = "Monthly", marker = list(color = bar_col),
             hovertemplate = hover_m) |>
    add_lines(y = avg, name = "12-month average", line = list(color = line_col, width = 3),
              hovertemplate = hover_m) |>
    layout(xaxis = list(title = "", rangeslider = list(visible = TRUE)),
           yaxis = list(title = "US$ million"),
           legend = list(orientation = "h", y = 1.12), hovermode = "x unified")
}

growth_chart <- function(df, monthly, rolling, col) {
  plot_ly(df, x = ~as.Date(month)) |>
    add_bars(y = monthly, name = "Month vs same month last year",
             marker = list(color = "#c8e6c9"),
             hovertemplate = "%{x|%b %Y}<br>%{y:.1f}%<extra>%{fullData.name}</extra>") |>
    add_lines(y = rolling, name = "Rolling 12 months vs a year earlier",
              line = list(color = col, width = 3),
              hovertemplate = "%{x|%b %Y}<br>%{y:.1f}%<extra>%{fullData.name}</extra>") |>
    layout(xaxis = list(title = ""), yaxis = list(title = "Growth, %", zeroline = TRUE),
           legend = list(orientation = "h", y = 1.12), hovermode = "x unified")
}

fy_tbl <- exp_m |>
  summarise(goods = sum(goods), services = sum(services), months = n(), .by = fy) |>
  arrange(fy) |>
  mutate(
    complete  = months == 12,
    fy_l      = str_c(fylab(fy), if_else(complete, "", " (to date)")),
    goods_g   = if_else(complete & lag(complete), goods / lag(goods) - 1, NA_real_),
    serv_g    = if_else(complete & lag(complete), services / lag(services) - 1, NA_real_),
    serv_sh   = services / (goods + services)
  )

fy_chart <- function(value, col, title) {
  plot_ly(fy_tbl, x = ~fy_l, y = value, type = "bar",
          marker = list(color = if_else(fy_tbl$complete, col, "#bdbdbd")),
          hovertemplate = "%{x}<br>US$ %{y:.2f} billion<extra></extra>") |>
    layout(title = list(text = title, x = 0), xaxis = list(title = "", categoryorder = "array",
           categoryarray = fy_tbl$fy_l), yaxis = list(title = "US$ billion"))
}
monthly_chart(exp_m, ~goods, ~goods12 / 12, "#a5d6a7", "#1b5e20")
growth_chart(filter(exp_m, !is.na(goods_yoy)), ~goods_yoy * 100, ~goods12_yoy * 100, "#1b5e20")
fy_chart(~goods / 1000, "#1b5e20", "Goods exports by fiscal year (grey = year to date)")
exp_m |>
  filter(fy >= cur_fy - 3) |>
  arrange(month) |>
  mutate(cum = cumsum(goods) / 1000, fy_l = fylab(fy),
         mname = month.abb[month(as.Date(month))], .by = fy) |>
  plot_ly(x = ~fy_month, y = ~cum, color = ~fy_l, type = "scatter", mode = "lines+markers",
          text = ~mname,
          hovertemplate = "%{text}: US$ %{y:.2f} billion<extra>%{fullData.name}</extra>") |>
  layout(xaxis = list(title = "", tickvals = 1:12, ticktext = month.abb[c(7:12, 1:6)]),
         yaxis = list(title = "US$ billion, cumulative"),
         legend = list(orientation = "h", y = 1.12))
if (svc_ok) cat("\n## Services exports\n")
monthly_chart(exp_m, ~services, ~serv12 / 12, "#b3e5fc", "#01579b")
growth_chart(filter(exp_m, !is.na(serv_yoy)), ~serv_yoy * 100, ~serv12_yoy * 100, "#01579b")
fy_chart(~services / 1000, "#01579b", "Services exports by fiscal year (grey = year to date)")
fy_tbl |>
  transmute(
    `Fiscal year` = fy_l,
    `Goods (US$ m)` = round(goods),
    `Goods growth` = scales::percent(goods_g, accuracy = 0.1),
    `Services (US$ m)` = round(services),
    `Services growth` = scales::percent(serv_g, accuracy = 0.1),
    `Services share of total` = scales::percent(serv_sh, accuracy = 0.1)
  ) |>
  select(where(\(x) !all(is.na(x)))) |>
  kable(align = "lrrrrr", caption = "Growth is shown only between complete fiscal years")
exp_m |>
  slice_tail(n = 12) |>
  transmute(
    Month = format(as.Date(month), "%b %Y"),
    `Goods (US$ m)` = round(goods),
    `Goods, vs a year earlier` = scales::percent(goods_yoy, accuracy = 0.1),
    `Services (US$ m)` = round(services),
    `Services, vs a year earlier` = scales::percent(serv_yoy, accuracy = 0.1)
  ) |>
  select(where(\(x) !all(is.na(x)))) |>
  kable(align = "lrrrr")
for (b in names(block_title)) {
  a <- A[[b]]
  if (is.null(a)) next
  cat("\n\n## Top 10 ", block_title[[b]], "\n\n", sep = "")
  cat(block_msg(a), "\n\n")
  print(plot_top(a))
  cat("\n\n")
  print(plot_trend(a))
  cat("\n\n")
  cat(table_top(a), sep = "\n")
  cat("\n\n")
}
cat("\n## Appendix: destination and product series\n\n")
series_info |>
  arrange(block, role, desc(mean_v)) |>
  transmute(Breakdown = block_title[block], Code = key, Name = label, Role = role,
            `Average monthly flow (US$ m)` = round(mean_v, 1)) |>
  kable() |>
  cat(sep = "\n")