Expertsystemen

 

Trend Finder

The idea behind Trend Finder is to highlight where unique values and combinations of values prevail per data point.

For example, in our demo NDA model (Non Disclosure Agreement) we could be interested in seeing location trends of our end users. First we type “location” into the search box, this will limit the data points to only ones containing the word “location” in them. Now when we select the data point “main.buyer.location” we get a chart showing what was recorded for that data point. It becomes clear to see that most of our end users indicated that the buyer was from New York.

Singular data point
Singular data point

If you select the ‘Short notation’ option, the legend labels will be value only, not ‘key = value’.

Short notation
Short notation

By clicking on the a legend items in the chart, we can remove it from the data, and refine the output to get the results we are interested in.

Hidden items
Hidden items

Never displayed is the count of cases where that data point was never offered, i.e. cases where the data point was in a node which was not in the path taken.

No Answer Given is the count of cases where the question was asked, but since the data point (interface) was not set to ‘required’ the users did not give an answer.

Sometimes you’ll want to omit the cases where the recorded value was ‘Never displayed’ or ‘No answer given’ - you can do this by checking the relevant box in the filters.

Trimmed
Trimmed

You’ll notice that directly above the output chart there is an ‘Omission ratio’ - this is to remind you that n% of the cases have been omitted owing to the filters.

What if we were also interested in knowing if there is a trend between the buyer location and the enterprise location. By selecting both data points, we can clearly see that there is a common trend where the buyer is from New York and the enterprise is in The Hague.

Combined data points
Combined data points

This kind of analysis can help you identify your audience and make calculated decisions based on that knowledge.

There is no limit to how many data points you can combine.