Visual Discovery
Last updated
Last updated
"Visual Discovery" is your ability to interrogate your data using collections of ADVIZOR charts. These features combine to provide Visual Discovery.
Focus: tell me more.
Touch a part of a display that represents data and dwell for a moment to learn more about it; this is called "focus". Only one item is in focus at once, and this focus item is shown in all charts that it is uniquely visible in.
For example, touch "Arco" to learn more about it. Since there is nothing in the second bar chart that represents only "Arco", no focus is shown in the bar chart.
If you want to focus immediately without dwelling for a moment, hold down the SHIFT key.
The information displayed for focus may be customized for some Charts via the property, available from the Property Explorer.
Automatic aggregation: slicing and dicing.
These two displays show Survey data aggregated (accumulated) in different ways: by Oil Company and by Fuel Grade. Both come from the same data source; the ADVIZOR charts compute the sums needed to display these different slices of the data from the same source automatically. You can dynamically change these as well.
The default aggregation is by "Count" the number of items in that category. You can also aggregate by summing another field in the data table, by taking the average, or by taking a ratio.
Notice that color stacking isn't supported for average aggregation; there is not an obvious mapping of the contribution of another field to the average.
Linked selection: what's the same?
Sweeping in a chart highlights (colors) the items you touched in that chart, and also highlights the same entries (however they appear) in any other charts. For example:
Sweep the bar "Regular" in the bottom bar chart on Fuel Grade. This selects the Fuel Grade "Regular".
Look at the top pie chart, which shows Oil Companies; the fraction of their fuel that is "Regular" is now highlighted.
In a pie chart, the selected subset can be interpreted as the fraction of that category with the selected value(s). So, you can see that oil companies sell most of their fuel in the "regular" grade (although the Fuel Grade bar chart showed you that as well), and that most fuel companies sell about the same fraction of their production as regular, with the notable exception of Chevron which has a lower fraction.
Touch Chevron with the mouse to focus on it and get more details.
By default a new selection replaces the current selection; you can modify the selection set instead by holding down the "SHIFT" key while dragging. This will toggle the selection state (selected becomes unselected, unselected becomes selected). Try this.
Exclusion: filter to an interesting subset.
Colors: use them to show data.
This example has colors mapped to "OilCo" (Oil Company). In the Pie Chart, which shows OilCo values directly, each pie slice is a single color. In the Bar Chart which shows Fuel Grades, each bar represents the sale of that fuel grade across all oil companies, so the colors are "stacked". The large blue portion of the "regular" bar corresponds to "Arco".
You can color by any field in your data and the color will be mapped onto all charts if possible. First, colors are taken from a "color scale", a linear sequence of colors. In this case, the scale used is the familiar rainbow ("Spectrum Bright") scale, but several others are available.
The same color scale is used for all charts.
Color is mapped onto a field by taking colors from a color scale based on the type of the field:
Categorical data: the data is ordered and colors are taken from the scale in order.
Continuous data: the range of the data is found and the colors from the color scale are mapped across that data range.
See also:
Besides selected and unselected, data may also be in the "excluded" state. When data is excluded, it is totally ignored: it does not appear in any charts and it is not used to compute ranges for axis. Data is excluded by applying exclusion to either the selected or unselected set, via the or buttons. Excluded data can be restored. Use exclusion to remove outliers that distort the display, or to focus on a subset of the data.