Information Display

The Charts and Graphs pattern set helps product managers, developers, and designers understand the advantages and disadvantages associated with visualizing data using the basic graph types. People use graphs to see large amounts of data at once, gain insight, and use that insight to take action. The choice of which graph to use depends on the types of data you want to display and the types of insight users are expected to be able to gain. If you are considering using multiple graphs, see the Guidelines for Using Multiple Graphs.

Use the following decision table to determine what pattern is appropriate based on user needs:

	Charts & Graphs	Analytic Grids	Analytic Field
Does the end user need to examine values to discern patterns or trends?	Yes	No	No
Does the end user need an easy way to look up a particular value?	No	Yes	No
Does the end user need only a single data point summary (for example, Up/Down)?	No	No	Yes

	Single-Quantity Graphs	Multi-Quantity Graphs	Category Graphs	Percentage Graphs	Total Graphs
Task: What data do you want to show or compare?	Single quantities with optional thresholds	Relationships between quantitative dimensions	Categories associated with simple quantities	Categories associated with percentages	Categories associated with totals

Key Concepts

User Types	Casual users are infrequent graph readers who can accurately interpret low-complexity graphs, such as bar graphs and pie charts. Experienced users are frequent graph readers who are comfortable reading higher-complexity graphs. Note that experienced users can also easily interpret graphs that are intended for casual users.
Data Dimension	A set of related data points of the same type and units, such as product names, years, costs, profits, and so on . Note that the term ”dimension” is used here to mean either quantitative or categorical data.
Correlation	Two quantitative data dimensions are correlated if they have a simple and recognizable relationship. For example, if profits rise with higher sales, the profit and sales dimensions have a positive correlation. If profits fall as expenses rise, the profit and expenses dimensions have a negative correlation. You use multi-quantity graphs to visualize correlations.
Trend	A trend is a correlation that can be used for future planning. For example, if profit and sales have a positive correlation, the trend is positive, indicating that future sales will likely result in future profits. Trends are emphasized by line graphs, a type of multi-quantity graphs; percent area graphs, a type of percentage graph; and stacked area graphs, a type of total graphs.
KPI	A Key Performance Indicator (KPI) is an aggregated single-quantity metric that is critical for the success of an organization, such as the performance of an operational, tactical, or strategic activity, or year-to-date expenses (see Single-Quantity graphs).

Data Dimension Types

Data Dimension Types example

	Categorical	Quantitative
	Categorical	Non-Sequential	Aggregate	Temporal
Data Values	Textual names (with no inherent numerical order)	Numeric quantities, possibly with duplicates	Ranges of numeric quantities	Time-based divisions
Examples	Products, geographic regions, customers	Salaries, prices	Salary ranges, price ranges	Years, quarters, weeks, seconds
Use in graphs	Bars, bar segments, pie slices, colors	Bar heights, pie areas, two-dimensional positions, gray values, bubble sizes	Bars, pie slices, colors	Horizontal axis
Relational terms	"items"	"data points," "variables"

Data dimensions are either categorical or quantitative.

Categorical data is the name of groups of things. Groups may be nested to form hierarchies. When shown in a graph, categorical data is typically shown using a different bar, pie slice, or color for each category or group.

Quantitative data resulting from counts or measurements are considered non-sequential because the data may have duplicate values. Plot non-sequential data using the heights of bars in bar graphs, the area of slices in pie charts, the two-dimensional positions of marks and bubbles in scatter graphs, and bubble graphs (respectively), and as gray values in gray scatter graphs and gray bubble graphs.

In contrast, aggregate and temporal data are sequential, do not have duplicate values, and are typically used to mark the scales of graphs. Aggregate data is formed when quantitative data is aggregated into uniformly sized bins or ranges of a sequential scale, such as when salary data is aggregated into salary ranges. Although aggregate data is quantitative, aggregated bins or ranges may also be considered a form of category or group. Aggregate data may, therefore, be shown using different bars or pie slices for each aggregate bin or range. Unlike categorical data, however, aggregate data has a well-defined quantitative order and should be shown with gray values instead of colors in scatter and bubble graphs.

Temporal data is uniformly spaced divisions of time such as years, weeks, and seconds. Due to established conventions, temporal dimensions should generally be plotted on a graph's horizontal axis.