By default, the Analysis page displays text mining results as Comments. You can switch between the different result views in the menu on the left.
Ordered from top to bottom, these buttons will take you to the different ways of viewing your results: Comments (default), Topics or Opinions. Each one offers a unique perspective of the results.
These three ways of viewing results can also be applied to Blacklists.
If you need a quick refresher on the difference between these concepts, revisit the article: How does Keatext analyze comments?.
The Comments view shows results in the natural language format, ranked by Record frequency. This view helps you understand the context quickly and easily. It is the default view.
Of the three views, this one offers the highest level of detail into what was expressed by the commenter. Each row represents a Comment group, which is a combination of one Topic (and all synonym variations), one Opinion (and all synonym variations) and the assigned sentiment category.
When selecting a Comment checkbox, the user has the options of hiding the Comment (sending to a Blacklist) or Tagging the Comment.
When you click on a Comment you are presented with the Comment details popup containing three tabs. The Records tab allows you to view all the Records that contain the selected Comment. Simply click the Record link to view the original record with all associated metadata. The Correlations tab allows you to create bar graphs that show the volume distribution of the selected Comment over a chosen metadata field. The Level of Interest graph shows the Record volume over time. At the bottom of the popup the Topic group and the Opinion group that the Comment belongs to are displayed. They can be selected and used as filters.
The Topics view helps monitor what is top of mind in the customer feedback. This helps prioritize what themes are most relevant. It shows the Topics in descending order of Record frequency, the distribution across the different sentiment categories, and a count of the number of Records it is present in.
Let's look at an example. In the screenshot below, the most frequent Topic is "app". This is what is top of mind for customers in the reviews we analyzed. It showed up mostly as Problems, with a similar amount of Praises and a relatively few Questions and Suggestions.
When you click on a Topic you are shown the Top Details, a similar popup we saw above on the Comments page.
The default way of viewing results is as a list. However, the user can choose to switch the Topics view to a chart.
The chart is a stacked bar that displays the proportion of sentiments expressed regarding each Topic. The darker the colour the higher the ratio, these are the topics you might want to concentrate on! Hover your mouse above the chart to get a full sentiment breakdown. You can easily export the Topics chart and add it to your presentation.
The Opinions view shows at a glance how commenters qualified the service or products they have given feedback on. This is a good view for brand management or users wishing to explore customer or employee emotions. The Opinions view ranks detected Opinions in order of Record descending frequency.
The Opinions view has the same features as the Topics view.