◆ Comparing record groups for visual analysis
Keshif's comparison selection offers rich ways to see trends across multiple data groups. This interaction is carefully designed to work seamlessly in all dashboards and across all charts, using effective, automated visualizations.
Understanding charts with compared data groups
Compared data groups are assigned a unique and perceptually distinguishable set of colors to visualize trends in aggregated charts, such as categorical bar charts and histograms, and in record charts, such as record maps or scatter plots. Keshif uses a consistent approach to compute data on record groups and visualize it using the same set of colors.
The resulting visualizations depend on data types, data analysis settings (such as breakdown mode) and visualization settings (such as visibility of whole aggregates, or the group-view as stacked or side-by-side).
Using Record Comparison Selection
- If side-by-side charts are active,
- Adjust the categorical bar heights to see multiple measurement labels for each compared group. The size multipliers are based on the supported number of comparisons, so if you have 3 comparisons, labels are shown on 3x or higher settings.
- Adjust the binning of interval charts to see more labels.
- If measurement labels are not visible because of small chart sizes, move the mouse over a compare selection breadcrumb on the top data status panel. This reveals the numbers specific to the desired record group.
- You can order categorical bar charts using trends in compared data groups. Learn how.
- You can compare up to five selections. Keshif uses distinguishable colors to show trends across different groups. Comparisons using too many colors is harder to perceive and analyze. This limit helps to ensure effective perception and analysis of data.
- You can only compare data groups within a single attribute. For example, you can compare people in different age groups, or different cities. You cannot compare an age group vs. a city group.
- Keshif disables highlight selections in data attributes that aren't being compared to ensure all selections are meaningful and comparable.
- You cannot make selections when using Adjust or Capture dashboard modes.