In this release we’re making an experimental feature available that we’ve been using internally for some time. It’s something we’ve gotten a lot of value out of and hope that it will empower you to discover more in your data!
Here’s what’s new...
There should be a new “Unmapped Verbatims” link in your sidebar. When you click on it, the first thing you’ll see is a large context network.
This network visualizes the top terms for verbatims in your data that have not been captured by a saved query. We find this is most useful when building a reporting framework -- as you construct a framework you can refer back to this screen and get a sense of the top narratives that you aren’t already tracking. You might find that there is something of interest to be added to your framework, or on the other hand validate that you’ve captured the key threads in your data.
Tip: Hover a term to see the frequency of unmapped verbatims it is found in and click through to the query screen. Note that on the query page you’ll see all instances of the term, whereas this screen only counts the number of unmapped verbatims that the term appears in.
If you scroll down the page you’ll notice some other widgets:
The “Most Frequent Unmapped Terms” chart displays the top terms present in unmapped verbatims ordered by frequency.; They should match the terms in the context network. You can click on an entry to go to the query page for that term. Again, note that on the query page you’ll see all instances of the term, whereas this screen only counts the number of unmapped verbatims it appears in.
The “Verbatims” widget simply shows all verbatims not captured by a saved query. It can be helpful in exploring the overall composition of your unmapped verbatims. For example, you might notice many instances of one-word verbatims like “Nothing” or “Good”.
Finally, the “Unmapped Data Statistics” shows a few key pieces of information about the unmapped verbatims in your data.
Most importantly % of all verbatims tells you how many of the verbatims in your data have not been captured by a saved query. This number will depend a lot on your data. For example, for a survey question like “Why did you give us this score?”, there may be a lot of respondents that answer “Good”. Tracking these answers may not give you a lot of insight, making it a reasonable decision to leave them as unmapped.
Additionally, you might find some terms in your data don’t appear frequently enough to warrant tracking, but collectively they comprise a modest portion of your data. While that’s interesting, tracking them in your reporting framework may not give you any extra insight. Having said that, we find that keeping this statistic around 20% or lower is a good reference point.
Avg words per verbatim gives you an indication of how long the unmapped verbatims in your dataset are. You could compare this to the Terms / Verbatim statistic on the “Summary” page, which, for example, might help you identify that the unmapped verbatims are significantly shorter than the rest of the verbatims in your data. Again, this might be indicative of a large degree of generic responses such as “Great” or “OK”.
That's it for this product update. If you have any thoughts or feedback of your own that you'd like to share with us, we'd love to hear from you!