Use the Deviations view to identify deviations of a facet value or facet for a given time period. The Deviations view focuses on how much the frequency of a facet value or facet deviates from the expected average for a given time period. You can use this view to observe seasonal patterns in your data or patterns that occur on a monthly or weekly basis.
After you select a facet from the Facet Navigation panel, the Deviations view displays a list of individual bar graphs, one for each value of the selected facet. Each individual bar graph displays deviation index scores to highlight facet values in your data that deviate from the normal distribution for the given time period. Deviation index scores measure the average occurrence of facet values on a selected date across the searched documents in the collection. Any frequency count that deviates (increases or decreases) from the computed average is highlighted. The highlighted deviations are shown in increasing shades of color, from yellow, which indicates a lower potential anomaly, to orange, to red. It is especially important to do further investigation on deviations that are highlighted in red.
For example, assume that red highlighting alerts you to the fact that the frequency count of fire-related incidents in 2002 and 2003 are much higher than the average frequency count that was computed over 10 years of data. To narrow the results so that you can discover the root cause of the problem, select the highlighted dates and add them to your search conditions.
You can select specific dates that you want to analyze more closely and show only those dates in the view. Select a bar in the chart and click Zoom in on selected dates to bring only those dates into focus. If the time scale is set to days or weeks, where the display of the vertical bars is dense, zooming in also makes it easier for you to identify and select the specific dates that you want to analyze. You can zoom out to reset the view and show frequencies for all dates according to the currently selected time scale.
To help you analyze results, you can select the charts that you want to focus on and view only those charts. You can also compare deviations for various facet values or subfacets by combining several charts into one.