Discover profiles a target group against the default set of descriptors. Harmoni ranks the rows in the Discover table so that the descriptors that most differentiate your target from others rise to the top, and those that are less likely to differentiate fall to the bottom of the table.
Learn how to use Discover to analyze your data and find items that define a target group.
In this article
- Getting Started with Discover
- Create a Discover Table
- Discover Analysis Features
- Add/Replace Descriptors or Selections
- Remove Descriptors from your analysis
- Switch back to a standard analysis
If you don't see the Discover drop zone in the analysis area, contact support@infotools.com to have the personalization key turned on.
1. Getting Started with Discover
The first step to creating a Discover analysis, is to have your descriptor variables setup. A Project Owner may set these up during the project creation and you can see them at the top of the project tree.
If they are not available in your project, you can create them yourself. Please review the Descriptor Manager article to learn how to setup Discover descriptors.
2. Create a Discover Table
With the descriptors created, there are a couple of ways to create a Discover analysis table:
a. From an existing analysis
From an analysis table, select the group you are interested in profiling. Selecting will enable the Discover option located in the modify menu.
- From the analysis table, select the target group
- Click Discover from the modify menu
- To return to the table, click the Back to Analysis icon
b. Drag and drop from the project tree into the analysis pane
From the project tree, select the group you are interested in profiling (i.e., Discover selection) and drag it into the Discover drop zone.
3. Discover Analysis Features
Change the Target
By default, discover uses the first element of the selected axis as the target. In the example below, the first element, Australia, becomes the target from the axis, Top 10 Markets. You can change the target by selecting the group you want to use and then choosing the option TARGET under Set in the modify menu.
- Drag the axis to profile into the Discover drop zone
- Harmoni creates a Discover analysis
- Select another element and click TARGET to change the target group
Changing the target does not change the statistical testing; it adjusts the ranking based on the newly selected target.
- Statistical testing uses the Total column, which is the average across all the groups. Each group is tested against the total.
- When ranking, the target group is compared to the rest of the groups (excluding the target).
Discover Ranking
A Discover table is ranked in two ways:
- Rows are ranked so that the descriptors that most differentiate the selected target come to the top of the table, and those that least differentiate the target fall to the bottom.
- Discover also sorts the groups across the table, so those most similar to the target are closest and those most different are furthest away.
Arrow Coloring
When ranking, Harmoni uses green arrows to indicate the descriptors that most differentiate the target and red arrows to indicate those that are less likely to differentiate the target.
For the coloring of the arrows, three shades of green and three shades of red add an additional level of differentiation within the level of likelihood. The darker the shade of green or red, the higher the probability.
The i-bar (information bar) at the bottom of the analysis screen includes the probability scale, along with the other details about the Discover analysis.
Information Bar
The information bar includes information that is important to know when you are reviewing a discover table. It automatically opens when you create a Discover analysis.
Percentage Base: Shows the axis percentages are based on
Reference: Shows the Total option being used as the reference in statistical testing
Target: The target group being assessed
Confidence Level: Confidence level for significance testing
Probability: Displays the probability and associated coloring for the arrows
Stat Testing: Bayesian Method (Probability testing used in Discover analysis)
3. Add/Replace Descriptors or Selections
After creating a Discover analysis, you can still interact with your discover descriptors and discover selections.
You can add or replace descriptors in your analysis with any axis from the project tree, even when not flagged as a descriptor. You can achieve this by dragging and dropping the relevant variables into the Discover Descriptors drop zone.
If you have multiple sets of discover descriptors, you can replace the whole group by dragging in another group and using Replace.
The same is true for your discover selections; you can add groups you are interested in profiling by dragging and dropping into the Discover Selections drop zone.
4. Remove Descriptors from your Analysis
You can remove descriptors from your discover analysis by using the Selector. Click on the interactive title and de-select the descriptors you don’t want.
5. Switch Back to a Standard Analysis
After creating a Discover analysis from a standard analysis, you can switch back to a standard analysis by using the Back to Analysis button in the modify menu.
This option is not available if you start a Discover analysis using the Discover drop zone.
Where to from here?
Learn more about Discover: