Need Help?

Overview: Navigating the Analytics Project Explorer ("What is everything on this screen?")

Last Updated: Apr 25, 2017
There are five key sections of the Analytics interface for exploring projects: 

1. Auto-Topic Suggestions (Topics List)
The first time you launch a project, this section automatically displays the 7 most important discussions, which each contains the 4 most important and frequent concepts within that discussion. You can keep or dismiss these suggestions -- and add your own.
Remember: a Topic = saved concept

2. Dynamic Concept Cloud (DCC)
This is the area in which you will do much of your initial exploration. The DCC looks similar to a word cloud, but is far more 
multi-dimensional, nuanced, and meaningful than a traditional word cloud. You can manipulate the concept cloud in several ways to analyze your data:
  • Instantly identify top related concepts by clicking on a term inside the DCC, which automatically generates a heat map of the term's related concepts.
  • Investigate a concept's most relevant related source documents (aka verbatims) in order to build your own unique Topic based on that concept.
  • Manipulate the DCC by selecting a term and dragging it to one corner of of the Cloud window. Your Cloud will reorganize itself and automatically display the most-related concepts in proximity to your selected term.
3. Top Related Concepts
Click the icon to expand the Concept Details pane (on the right). In this section, you can:
  • Investigate any concept's Top Related Concepts (TRC) by searching for or selecting a topic and scrolling through the TRC list. The top 25 related concepts display.
  • By hovering over a TRC, instantly derive hard metrics (number of document matches and percentage of documents in which the concept appears).
4. Relevant Documents (Verbatims)
Select a concept to explore to display its corresponding source documents ranked in order of most to least relevant (50 display in the list). You can download more by clicking Download Matching Documents. 

If document URLs were included in your data set, you can also view the document in its original form (tweet, Yelp or Amazon review, etc.) by clicking on the underlined document title.

5. Analysis Exports

Click Analysis Exports at the top to access these exportable metrics as spreadsheets that include numbers and charts to help visualize the information for reporting (see page 6). The match counts exports contain the numbers and percentages of both exact and conceptual matches contained in all documents and within each subset.

Download the:
  • Top Match Counts to quantify a Top Concept’s overall frequency and importance in a dataset.
  • Topic Match Counts to quantify a saved Topic’s overall frequency and importance in a dataset.
  • Topic Timelines to measure association scores over time (source documents must be assigned dates).
  • Topic-Topic Association Scores to compare a topic’s association score with any other topic (see NOTE).
  • Topic-Subset Association Scores to measure a saved topic’s relative prevalence within a subset.

Association Scores above:

  • 0.7 are particularly strong
  • 0.2 are notable

Often the most interesting related concepts fall between the 0.2 – 0.7 range 
Topic-Topic Associations scores have a different threshold (0.3 and above are strong associations)

6. Subsets 
Upon opening a new project, the default "All documents" view will display. However, if you wish to select a particular document subset to explore, click All Documents to display a drop-down list of all the subsets within your data set, and then click the subset you wish to explore. You will see the total number of documents in parentheses next to each subset name. (Document subsets are created in the source data before uploading a project.)

You can also search for concepts (top middle), use the axes overlay feature and reset the DCC by using the controls at bottom/middle of your screen.

Consult the QuickStart Guide and other documentation (such as the Theme Identification guide) for workflow steps and analysis techniques.


More Support
seconds ago
a minute ago
minutes ago
an hour ago
hours ago
a day ago
days ago
Invalid characters found