Pareto Chart
Bar chart sorted by descending value with a cumulative-percentage line — the classic "vital few" tool.
Anatomy of a Pareto Chart
The Pareto chart is one of the "7 QC tools" and combines a bar chart sorted by descending value with a cumulative-percentage line. It makes the 80/20 principle ("vital few — trivial many") visually obvious: a few categories drive most of the problem.
Sorting: Bars are always sorted by descending value. At a glance you see which categories contribute the most — unlike a plain bar chart, where the X axis stays in input order.
Cumulative percentage line: A line over the bars shows, for every point, the cumulative share (in %) of the grand total. If it reaches the 80 % mark after three categories, those three together account for 80 % of the problem.
80 % threshold: A horizontal reference line at 80 % marks the classic Pareto boundary. It helps to read off the "vital few" — all bars to the left of the intersection belong to them. The chart editor lets you change the threshold or hide it (value 0).
"Other" bucket: With many categories, the long tail (smallest contributions) is collapsed into a single "Other" bar at the end. The chart stays readable without losing the tail. The chart editor lets you tune the maximum bar count.
Classic Six-Sigma uses: prioritising defect types, weighting complaint reasons, identifying cost drivers. The module accepts a categorical column (frequency mode), optionally a value column (sum per category — e.g. complaint cost) and/or a grouping column for stacked Pareto charts.
Pitfalls
Frequency ≠ importance: A frequently-occurring defect type is not necessarily the most expensive one. When each complaint carries a different cost, fill the value column (Y) and sort by sum — otherwise the chart points the focus to the most frequent, not the most damaging issues.
Too many categories: If all contributions are similar, there are no "vital few" — the Pareto principle does not apply. A Pareto chart with 20 nearly equal bars is more a symptom of a definition problem: the categories may be too granular or too generic.
Wrong aggregation: When the Y column is set, you must choose whether to sum or to average across the category. For costs/damage/quantities, "sum" is right (total contribution); for an average complaint amount, use "mean". The wrong choice produces the wrong ranking.
Comparing over time: Pareto charts are snapshots — the bar order changes between periods. For trends, prefer line or stacked bar charts and use a Pareto chart only as a per-period snapshot.
"Other" bucket too large: If the "Other" bar is itself among the largest, the top-N cut is set too tightly. Increase the max bar count in the chart editor until "Other" is smaller than the individual top categories.
Examples
This module ships with the following example datasets — load any of them in the app with a single click.