pbPassingBI
/
Building reports beginner 6 min

Choosing and building charts

Time series, bar, pie, scatter and geo — and which question each answers.

What you'll be able to do
  • Match chart type to question
  • Configure the common charts
  • Avoid the chart types that mislead

The mapping

QuestionChart
How has this changed over time?Time series
How do categories compare?Bar
Where is this happening?Geo map
Are two metrics related?Scatter
What is the composition?Stacked bar — rarely pie
How do users move through?Funnel or Sankey (community)

Time series

Needs a date dimension and at least one metric. A second metric adds a line; a breakdown dimension splits into one line per value.

Breakdown dimension cardinality

A breakdown on something with 30 values produces 30 lines and an unreadable chart. Add a filter to the top few, or use a table instead.

Bar charts

Sort descending by the metric — the ranking is normally the point. Use horizontal bars when the category labels are long, which is most of the time with campaign or page-path names.

Stacked bars work for composition, but comparing the middle segments across bars is genuinely hard because they share no baseline.

Pie charts

Available, and usually the wrong choice. People compare angles poorly, and anything past three or four slices is read more accurately as a sorted bar chart.

Geo maps

The geo field needs a recognised type — country, region, city, or latitude/longitude. Set that in the data source.

A filled map shades regions by a metric. Bear in mind it encodes area as much as value, so large sparse regions dominate visually; a bubble map is often more honest.

Optional metrics

In Setup, Optional metrics lets the viewer add metrics themselves from a menu. It keeps the default chart clean while letting people who want more detail get it — a much better answer than adding six metrics by default.

Key points
  • Sort bar charts by metric and use horizontal bars for long labels
  • A breakdown dimension with many values makes a time series unreadable
  • Optional metrics keep a chart clean while allowing depth on demand
Check yourself