# Breakdown: Absolute vs. Relative Breakdown Charts

You can discover new trends in compared data groups by visualizing them using relative or absolute breakdown charts.

• The absolute-breakdown charts show the absolute size of compared data groups (such as count of records in groups).
• The relative-breakdown charts show the relative (percentage) frequency of compared data groups within specific data aggregates (such as categories).

You can switch the mode with just a single click, and use stacked or side-by-side charts with both breakdown modes. All aggregate charts in a dashboard use the same breakdown mode.

#### How can you set breakdown mode to absolute or relative?

• Click on the chart axis, or
• Use the dashboards config panel, via the gear  icon on the upper-left dashboard corner.

#### Example

Breakdown comparison analysis of four quadrants on a homicides data set.

Analysis of homicides on hispanic race in NW quadrant:

Absolute mode: Among the 96 total cases for hispanic group, 76 were in NW (violet color). The bars for hispanic group are tiny because of the few number of homicides.

Relative mode: 79% (76/96) of the hispanic cases are in NW. The relative quadrant breakdown makes it easier to observe trends by race and quadrant.

Absolute mode: Time-chart emphasizes the downward trend of homicide cases over time. In 2009, the total number dropped below 150 for the first time since 2000.

Relative mode: Time-chart reveals the relative number of homicides per four quadrants over time (measured annually). In 2009, NW (violet color) had 24% of the incidents, compared to 28% of the incidents in the previous year.

#### When can you analyze data with relative breakdown?

Keshif helps you prevent misleading data analysis when relative breakdowns would be inaccurate or inappropriate. Specifically,

• The aggregation function should not be "average".
Example: You cannot analyze relative "average-age" by gender.
• The aggregation function should not be "total/sum" with a metric that includes negative values.
Example: You cannot analyze relative "total-temperature" by season, as temperature may be negative.
• The dashboard should have at least one compared selection.

When you remove compared selections, the relative breakdown will be auto-disabled.