Insights in visualizations
Insights in visualizations provide analytic insights that can help users to detect and validate any important relationships and meaningful differences based on the data that is presented by the visualization.
Insights are controlled by and summarized in the Insights box available in every eligible visualization. When you turn on insights, the summary appears in the Insights box and related visualization elements are highlighted. Details are provided in the corresponding tooltip message. You can control each available insight separately.
The type of insights depends on the displayed data by the visualization. Available types of insights are Average, Predictive strength, Meaningful Differences, Fit line, and Most Frequent. Average provides the mean of the displayed summaries, and most frequent the category or category that appears most often in the data. The rest of the insights depend on more advanced analytics and statistical tests. The goal is to provide reliable information that you can use for an enhanced description of the viewed data and discovery of any relationships that are expected to be found in the population that is represented by this data.
Insights analysis is always based on the same data rows that are used to create the summaries displayed in the visualization. This means that full data is used for insights unless any filtering is applied to the original data.
Some statistical tests and analytics that are used in insights require not only the data summaries that are displayed in the visualization, but also some additional summarizations. For example, test of meaningful differences across multiple categories of an explanatory field requires counts and variances for each category in addition to the displayed data. These additional summaries are obtained from a database together with the summaries that are needed for the visualization. All summaries are processed by the insights but only the required summaries are available in the visualization. Insights analysis is always based on the same data rows that are used to create the summaries displayed in the visualization.
If insights are not immediately available in a visualization, one of the following reasons might apply:
- The visualization type itself does not support insights.
- The data in the visualization may have been clipped.
- The combination of summarization level, field type, and field role of a selected field does not match the requirements for any of the available insights.
- Supported visualization types for insights
The following visualization types support insights:
- Heat map
- Hierarchy bubble
- Line and column
- Packed bubble
- Stacked bar
- Stacked column
- Tree map
- Word cloud
Small multiple extensions are supported for some insights including the Most Frequent and Meaningful Differences.
- Summarization levels
Supported summarization levels are Count, Average, Sum, Minimum, and Maximum. Any other values such as Count distinct might prevent the insights from being suggested. Certain algorithms support only specific summarization levels. Changing the default summarization level to one of the supported values might potentially help with enabling insights.
- Field types
Field types can be internally designated as continuous or categorical depending upon the values of the selected field.
Field type Description Categorical A variable that can take on one of a limited, and usually fixed, number of possible values. A categorical variable assigns each individual or other unit of observation to a particular group or nominal category based on some qualitative property. For example, the country a person lives in. Continuous A variable that is used to describe numeric values, such as a range of 0-100 or 0.75 - 1.25. A continuous value can be an integer, real number, or date and time.
- Field roles
IBM® Cognos Analytics assigns a role to each of the field slots in a supported visualization. A field role might be designated as one of the following depending on the slot of the visualization.
Field role Description Response A variable that is predicted and can also be referred to as the target or dependent variable. It is commonly on the Y-axis. Explanatory A variable which helps to explain changes in the response and is also referred to as the predictor or independent variable. It is commonly on the X-axis. Group A variable that is treated as explanatory or an optional grouped factor which helps to determine the number of models built in the algorithm. For example, this can correspond to the Color slot of a Column visualization. Weight A variable which defines the optional regression weights, which are used to calculate the regression model. For example, this can correspond to the Size slot of a Bubble visualization. Repeat A variable which creates small multiples, with the visualization repeated once for each distinct value of the variable. For example, this can correspond to the Repeat (rows) slot of a pie visualization. Points A variable which defines the shaping of the data and data points used to calculate the model. For example, this can correspond to the Points slot of a Scatter visualization.
As a general example, in a bar visualization with the following slots, the role mappings in this visualization are defined as:
- Bars (y-axis), explanatory
- Length (x-axis), response
- Color, group