Infotools Harmoni Discover lets you find stories in data faster by profiling the groups that matter and comparing them with others.
Discover is a tool to profile a target group against a set of descriptor variables. It will show you which of these descriptors best describe your group and compare these with other groups.
For example, Discover will give you a profile of, say, those who prefer your brand and then compare that profile with the same for all the other brands. The profile (or descriptors) you use to compare the brands might be demographic, psychographic, behavioral or some combination of all these.
Discover will present a table view with the variables that most distinguish your chosen group from the rest at the top. It also sorts the groups across the table so that those (brands, say) that are most similar to the one you’ve chosen are closest and the ones that are most different are furthest away.
How does discover work?
Discover works by performing a series of statistical tests on each descriptor, comparing the cell value for each group with the value for those, not in the group (the rest). These values are compared using Bayesian Statistics to calculate the probability that the group value is greater (or lower) than the rest values.
In the table, the descriptor values are ranked by the probabilities of the cells in the target group (which will be shown in the first column). The descriptors for which the target group has the highest probabilities of being greater than the Rest appear at the top of the table. Similarly, those descriptors for which the target group has the lowest probabilities of being higher than the Rest will appear at the bottom of the table. (Descriptors with average probabilities of being greater appear mid-way down the table, but these will be suppressed for tables with a large number of descriptors.)
Discover shows the target group in the first column of the table and the rows of the table are sorted based on this target group. The other columns (groups) are sorted by order of similarity to the target column. For example, if your groups are brands, your target brand will be in the first column (next to the Total column that is), the brand next to it will be the one with the most similar profile according to the descriptors down the left of the table. The brand furthest away from your brand will be the least similar. The similarity is calculated using Robinson's agreement metric between the target and comparison columns.