Executive Summary

The 80/20 split across 30 categories

Analysis Overview

Pareto analysis of 30 Defect Type categories across 506 rows.

N Categories30

N Rows506

Vital Few7

A Class7

Interpretation

The Pareto principle identifies where to focus effort to move the total. This analysis ranks 30 Defect Type categories by total Cost across 506 records, totalling 27,631. The vital few—just 7 categories (23.3% of the set)—carry 83.1% of the total Cost. These A-class categories are the leverage points: improving them yields proportionally larger gains than spreading effort across all 30. The remaining 23 categories contribute only 16.9%, making them lower-priority unless their defects are easily corrected. The concentration (Gini: 0.729) confirms that cost is highly skewed toward a small subset, validating the focus on the top 7.

Data Preparation

Data Quality

Rows, categories, negatives handling, and the chart rollup.

Initial Rows521

Final Rows506

Rows Removed15

N Categories30

Negative Dropped15

Zero Kept5

Interpretation

Data quality is sound for Pareto analysis. Of 521 rows loaded, 506 were retained across 30 Defect Type categories, with a grand total Cost of 27,631. Fifteen records with negative Cost values (totalling -768) were dropped because Pareto assumes non-negative contributions; five zero-value records were retained. The chart displays the top 20 categories individually and rolls the remaining 10 into an Other bar (371.49 in value), but all 30 categories are included in the vital-few and ABC class calculations. This approach preserves analytical integrity while keeping the visual uncluttered.

Visualization

Pareto Chart

Ranked category totals with the cumulative share.

Interpretation

The ranked bars fall steeply: D01 leads at 6,402.57 (23.17%), followed by D02 at 4,759.71 (17.23%) and D03 at 3,925.18 (14.21%). The cumulative share climbs sharply through the first seven categories, reaching 83.1% at D07 (1,137.56), then flattens dramatically. By D10 (625.34), the cumulative is already 91.41%, and the tail from D11 onward each contributes less than 1.5% individually. This steep-then-flat profile is the signature of Pareto dominance: a handful of categories account for most cost, while the rest scatter thinly. The visual confirms that effort on the top 7 yields disproportionate impact.

Data Table

ABC Classes

Category counts and value share for the A, B, and C classes.

Class	Categories	Value	Value Share PCT	Description
A	7	22,962	83.1	Vital few - the top categories that together reach about 80 percent of the total.
B	7	3,467	12.5	Useful many - the next categories carrying the 80 to 95 percent band.
C	16	1,202	4.4	Trivial many - the long tail below the last 5 percent.

Interpretation

Class A (7 categories) holds 22,962 in Cost—83.1% of the total—and is the vital few demanding tight management and immediate intervention. Class B (7 categories) carries 3,467 (12.5%), spanning the 80–95% cumulative band; these merit batched or periodic review. Class C (16 categories) totals 1,202 (4.4%), the trivial many below 95%; these can be handled in bulk or deprioritized. The three-tier split clarifies governance: A categories require individual scrutiny and targeted cost-reduction initiatives; B categories warrant consolidated monitoring; C categories absorb minimal resources. This segmentation aligns effort intensity with value impact.

Data Table

Top Contributors

The highest-value categories ranked, with cumulative share.

Rank	Category	Value	Share PCT	Cumulative PCT
1	D01	6,403	23.17	23.17
2	D02	4,760	17.23	40.4
3	D03	3,925	14.21	54.6
4	D04	2,886	10.45	65.05
5	D05	2,253	8.15	73.2
6	D06	1,598	5.78	78.99
7	D07	1,138	4.12	83.1
8	D09	863	3.13	86.23
9	D08	805	2.91	89.14
10	D10	625	2.26	91.41
11	D11	391	1.42	92.82
12	D12	297	1.07	93.9
13	D14	274	0.99	94.89
14	D15	210	0.76	95.65
15	D13	197	0.71	96.36

Interpretation

D01 dominates at 6,403 (23.17% share), followed by D02 at 4,760 (17.23%) and D03 at 3,925 (14.21%). The cumulative share crosses the 80% vital-few threshold at rank 7 (D07: 1,138, 4.12% share, cumulative 83.1%). Below rank 7, individual category shares drop sharply—D09 is 3.13%, D08 is 2.91%, and by rank 10 (D10) the share falls to 2.26%. This ranked list is the concrete priority roadmap: the first seven categories are the cost-drivers to address; the remainder are secondary. Focusing resources on reducing D01, D02, and D03 alone would cut total cost by more than half (54.6%).

Methodology

Statistical methodology and diagnostics for Pareto Analysis — The Vital Few (80/20)

Statistical Method

Pareto Analysis — The Vital Few (80/20)

Standard-library analysis: find the few categories that drive most of the total, straight from a raw category/value table. The analysis totals your value by category, ranks them biggest to smallest, builds the cumulative 80/20 curve, and reports the vital few that reach 80%, the ABC classes (A/B/C), and how concentrated the total is — the workhorse Lean and Six Sigma tool for defects, complaints, revenue concentration, or SKUs. Map one category column and one numeric value column.

Data

N = 506 observations · 15 excluded

Assumptions

Values are contributions to a total (counts or amounts), not rates or ratios
Negative values are refunds or corrections and are dropped; zeros are kept
Categories that repeat across rows are summed before ranking

Limitations

Pareto ranks and concentrates a total — it does not test significance or explain why a category is large
A very long tail of tiny categories can make the 80/20 split look sharper than day-to-day operations feel
The chart caps at the top 20 categories (the rest pool into Other), though every category still counts in the vital-few and ABC totals

Software & Citation

MCP Analytics · mcpanalytics.ai

Code Appendix

Analysis Code

Complete R source code for this analysis

Pareto Analysis — The Vital Few (80/20)

Finds the few categories that drive most of the total. Aggregates a raw category/value table, sorts the categories from biggest to smallest, builds the cumulative curve, and reports the "vital few" that reach 80% of the total, the ABC classes, and how concentrated the total is.

Why This Method?

The Pareto principle — roughly 80% of effects come from 20% of causes — is the workhorse of Lean and Six Sigma. It turns a long, flat list of defects, complaints, SKUs, or revenue lines into a ranked priority list: fix the vital few and you move most of the total. It needs no model, only a category and an amount.

What This Analysis Covers

The vital few: the smallest set of categories reaching 80% of the total
ABC classes (A up to 80%, B 80-95%, C the tail)
The Pareto chart (ranked bars with the cumulative share)
Concentration: the top category and top-three shares

Standard Library

Platform standard-library module (LAT-1441): runs on ANY dataset via the semantic mapping {category, value}. All narrative is derived from the user's own column names and computed values.

suppressPackageStartupMessages(library(DT))
suppressPackageStartupMessages(library(htmlwidgets))
suppressPackageStartupMessages(library(arrow))
suppressPackageStartupMessages(library(knitr))
suppressPackageStartupMessages(library(rmarkdown))
suppressPackageStartupMessages(library(dplyr))
suppressPackageStartupMessages(library(tidyr))
suppressPackageStartupMessages(library(ggplot2))
suppressPackageStartupMessages(library(stringr))
suppressPackageStartupMessages(library(lubridate))
suppressPackageStartupMessages(library(broom))
suppressPackageStartupMessages(library(Matrix))
suppressPackageStartupMessages(library(cluster))
suppressPackageStartupMessages(library(data.table))

Core Analysis Pipeline

Number formatter — plain rounded strings with thousands separators.

Table numerics are formatted through this so large totals never serialize in scientific notation (e.g. 2.76e+04) inside a rendered table.

fmt_num <- function(x, digits = 0) {
  format(round(as.numeric(x), digits), big.mark = ",", scientific = FALSE,
         trim = TRUE, nsmall = digits)
}

Gini concentration (0 = every category equal, 1 = one holds it all).

Defensive: needs a positive total and >=2 categories, else returns NA.

gini_concentration <- function(v) {
  v <- v[!is.na(v)]
  n <- length(v)
  s <- sum(v)
  if (n < 2 || s <= 0) return(NA_real_)
  vs <- sort(v)
  (2 * sum(seq_len(n) * vs) / (n * s)) - (n + 1) / n
}

compute_shared <- function(df, params, col_map = list()) {
  # === SHARED EXPORTS ===
  #   initial_rows/final_rows/rows_removed   $ row accounting
  #   n_missing_cat / n_missing_value / n_negative / n_zero  $ preprocessing counts
  #   negative_total          $ numeric — summed value of dropped negatives
  #   cat_name / value_name   $ humanized user column names
  #   agg_df                  $ data.frame(category, value, share_pct,
  #                             cumulative, cumulative_pct, rank, abc) sorted desc
  #   n_categories / total_value                $ integer / numeric
  #   vital_few_n / vital_few_cat_share / vital_few_value_share
  #   top_category / top_share / top3_share     $ concentration headline
  #   gini                    $ numeric — Gini concentration
  #   abc_summary_df          $ data.frame(class, categories, value,
  #                             value_share_pct, description)
  #   n_a / n_b / n_c         $ integer — ABC class sizes
  #   pareto_df               $ data.frame(category, value, cumulative_pct) top 20 + Other
  #   top_contributors_df     $ data.frame(rank, category, value, share_pct, cumulative_pct)
  #   metrics / json_output
  # === /SHARED EXPORTS ===

Step 1: Required semantic columns + humanized names

initial_rows <- nrow(df)
  cat_name   <- humanize_semantic("category", col_map)[1]
  value_name <- humanize_semantic("value", col_map)[1]
  for (need in c("category", "value")) {
    if (!need %in% names(df)) {
      stop(sprintf("Required column &#x27;%s' is not mapped.",
                   humanize_semantic(need, col_map)[1]))
    }
  }

Step 2: Category — character, drop blank/missing

cat_v <- trimws(as.character(df$category))
  keep_cat <- !is.na(cat_v) & cat_v != ""
  n_missing_cat <- sum(!keep_cat)
  cat_v <- cat_v[keep_cat]
  val <- df$value[keep_cat]

Step 3: Value — 95% numeric coercion rule; drop NA

if (!is.numeric(val)) {
    conv <- suppressWarnings(as.numeric(as.character(val)))
    n_orig <- sum(!is.na(val) & trimws(as.character(val)) != "")
    if (n_orig == 0 || sum(!is.na(conv)) < 0.95 * n_orig) {
      stop(sprintf(
        "Column &#x27;%s' is not numeric enough to total up — fewer than 95%% of its values could be read as numbers.",
        value_name))
    }
    val <- conv
  }
  keep_val <- !is.na(val)
  n_missing_value <- sum(!keep_val)
  cat_v <- cat_v[keep_val]
  val <- val[keep_val]

Step 4: Drop negatives (Pareto assumes non-negative contributions),

keep zeros. Count and total the dropped negatives for the narrative.

neg_mask <- val < 0
  n_negative <- sum(neg_mask)
  negative_total <- if (n_negative > 0) sum(val[neg_mask]) else 0
  cat_v <- cat_v[!neg_mask]
  val <- val[!neg_mask]
  n_zero <- sum(val == 0)

  final_rows <- length(val)
  rows_removed <- initial_rows - final_rows
  if (final_rows < 5) {
    stop(sprintf(
      "Only %d usable rows remain after cleaning &#x27;%s' and '%s' — need at least 5.",
      final_rows, cat_name, value_name))
  }

Step 5: Aggregate (sum) by category, sort biggest to smallest

agg <- data.frame(category = cat_v, value = val, stringsAsFactors = FALSE) %>%
    group_by(category) %>%
    summarise(value = sum(value), .groups = "drop") %>%
    as.data.frame(stringsAsFactors = FALSE)
  agg <- agg[order(-agg$value), , drop = FALSE]
  rownames(agg) <- NULL

  n_categories <- nrow(agg)
  total_value <- sum(agg$value)
  if (total_value <= 0) {
    stop(sprintf(
      "The total of &#x27;%s' across all categories is zero, so there is no share to distribute.",
      value_name))
  }

Step 6: Shares + cumulative curve

agg$share_pct      <- round(100 * agg$value / total_value, 2)
  agg$cumulative     <- cumsum(agg$value)
  agg$cumulative_pct <- round(100 * agg$cumulative / total_value, 2)
  agg$rank           <- seq_len(n_categories)

Step 7: The vital few — smallest prefix reaching 80% cumulative.

cumulative_pct is monotonic and its last value is 100, so a match always exists; which.max returns the first TRUE (guarded fallback regardless).

reach80 <- agg$cumulative_pct >= 80
  vital_few_n <- if (any(reach80)) which.max(reach80) else n_categories
  vital_few_cat_share   <- round(100 * vital_few_n / n_categories, 1)
  vital_few_value_share <- round(agg$cumulative_pct[vital_few_n], 1)

Step 8: ABC classes (A up to 80%, B 80-95%, C the tail)

reach95 <- agg$cumulative_pct >= 95
  b_end <- if (any(reach95)) max(which.max(reach95), vital_few_n) else n_categories
  abc <- rep("C", n_categories)
  abc[seq_len(vital_few_n)] <- "A"
  if (b_end > vital_few_n) abc[(vital_few_n + 1):b_end] <- "B"
  agg$abc <- abc

  n_a <- sum(abc == "A"); n_b <- sum(abc == "B"); n_c <- sum(abc == "C")

  abc_desc <- c(
    A = "Vital few - the top categories that together reach about 80 percent of the total.",
    B = "Useful many - the next categories carrying the 80 to 95 percent band.",
    C = "Trivial many - the long tail below the last 5 percent."
  )
  abc_summary_df <- do.call(rbind, lapply(c("A", "B", "C"), function(cl) {
    idx <- abc == cl
    data.frame(
      class = cl,
      categories = sum(idx),
      value = fmt_num(sum(agg$value[idx]), 0),
      value_share_pct = round(100 * sum(agg$value[idx]) / total_value, 1),
      description = unname(abc_desc[cl]),
      stringsAsFactors = FALSE
    )
  }))
  rownames(abc_summary_df) <- NULL

Step 9: Concentration headline

top_category <- agg$category[1]
  top_share    <- agg$share_pct[1]
  top3_n       <- min(3, n_categories)
  top3_share   <- round(sum(agg$share_pct[seq_len(top3_n)]), 1)
  gini         <- round(gini_concentration(agg$value), 3)

Step 10: Pareto chart data — top 20 categories, rest rolled into

"Other" (chart only; the vital-few and ABC counts use the FULL data).

cap <- 20
  if (n_categories > cap) {
    head_df <- agg[seq_len(cap), c("category", "value", "cumulative_pct")]
    other_value <- total_value - sum(agg$value[seq_len(cap)])
    pareto_df <- rbind(
      head_df,
      data.frame(category = "Other", value = round(other_value, 2),
                 cumulative_pct = 100, stringsAsFactors = FALSE)
    )
  } else {
    pareto_df <- agg[, c("category", "value", "cumulative_pct")]
  }
  pareto_df$value <- round(pareto_df$value, 2)
  rownames(pareto_df) <- NULL

Step 11: Top contributors table — top 15, values as plain strings

tc_n <- min(15, n_categories)
  top_contributors_df <- data.frame(
    rank = agg$rank[seq_len(tc_n)],
    category = agg$category[seq_len(tc_n)],
    value = fmt_num(agg$value[seq_len(tc_n)], 0),
    share_pct = agg$share_pct[seq_len(tc_n)],
    cumulative_pct = agg$cumulative_pct[seq_len(tc_n)],
    stringsAsFactors = FALSE
  )
  rownames(top_contributors_df) <- NULL

  metrics <- list(
    `Categories`            = n_categories,
    `Vital Few(80%)`       = vital_few_n,
    `Vital Few Share %`     = vital_few_cat_share,
    `Top Category Share %`  = top_share,
    `Top 3 Share %`         = top3_share,
    `Concentration(Gini)`  = gini
  )

  json_output <- list(
    answer = paste0(
      "Pareto analysis of ", format(n_categories, big.mark = ","),
      " categories from ", format(final_rows, big.mark = ","), " rows: ",
      "the vital few is ", vital_few_n, " ", cat_name,
      if (vital_few_n == 1) " category" else " categories", " (",
      vital_few_cat_share, "% of them) driving ", vital_few_value_share,
      "% of total ", value_name, ". The single biggest is ", top_category,
      " at ", top_share, "%; the top ", top3_n, " together hold ", top3_share,
      "%. ABC split: A = ", n_a, ", B = ", n_b, ", C = ", n_c, " categories."
    ),
    cards = lapply(
      c("tldr", "overview", "preprocessing", "pareto_chart",
        "abc_classes", "top_contributors"),
      function(cid) list(id = cid, metrics = metrics)
    )
  )

  list(
    initial_rows = initial_rows, final_rows = final_rows,
    rows_removed = rows_removed,
    n_missing_cat = n_missing_cat, n_missing_value = n_missing_value,
    n_negative = n_negative, negative_total = negative_total, n_zero = n_zero,
    cat_name = cat_name, value_name = value_name,
    agg_df = agg, n_categories = n_categories, total_value = total_value,
    vital_few_n = vital_few_n, vital_few_cat_share = vital_few_cat_share,
    vital_few_value_share = vital_few_value_share,
    top_category = top_category, top_share = top_share, top3_share = top3_share,
    top3_n = top3_n, gini = gini,
    abc_summary_df = abc_summary_df, n_a = n_a, n_b = n_b, n_c = n_c,
    pareto_df = pareto_df, top_contributors_df = top_contributors_df,
    metrics = metrics, json_output = json_output
  )
}

Your data has more stories to tell. Run any analysis on your own data — 60+ validated R modules, interactive reports, AI insights, and PDF export. 500 free credits on signup.

Try Free — No Signup Sign Up Free

Something wrong? Report it and we'll fix it — or get your credits back

Download PDF · Print · Share

What's wrong with this card?

Executive Summary

Analysis Overview

Data Quality

Pareto Chart

ABC Classes

Top Contributors

Methodology

Statistical Method

Analysis Code

Pareto Analysis — The Vital Few (80/20)

Why This Method?

What This Analysis Covers

Standard Library

Core Analysis Pipeline

Number formatter — plain rounded strings with thousands separators.

Gini concentration (0 = every category equal, 1 = one holds it all).

Step 1: Required semantic columns + humanized names

Step 2: Category — character, drop blank/missing

Step 3: Value — 95% numeric coercion rule; drop NA

Step 4: Drop negatives (Pareto assumes non-negative contributions),

Step 5: Aggregate (sum) by category, sort biggest to smallest

Step 6: Shares + cumulative curve

Step 7: The vital few — smallest prefix reaching 80% cumulative.

Step 8: ABC classes (A up to 80%, B 80-95%, C the tail)

Step 9: Concentration headline

Step 10: Pareto chart data — top 20 categories, rest rolled into

Step 11: Top contributors table — top 15, values as plain strings

Report an Issue