A Record is a single row in a dataset (CSV file used for import). Records are comprised of both the long form qualitative textual data for analysis and the accompanying metadata, per commenter. Since a single record may contain many comments, it can be useful to access the original records in full for deeper context while data mining – what people are saying as a whole, rather than in an isolated comment. You can drill right down to the original Records in Focus Recommendations, Dashboards and on the Analysis page.
Keatext's analysis summarizes your data and displays the results in a clear, concise manner by producing a list of Comment groups - Comments containing the same Topic and Opinion - in descending order of frequency in the Comments view on the Analytics page. Displaying information in a natural language format helps you understand the context quickly and easily.
Viewing a Record
On the Analysis page, you can pull up a Comment details popup by clicking on any comment in the list, such as “delivery fee is outrageous".
The Comment details popup contains 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.
Selecting “View record” allows you to see the entire Record, including all associated metadata.
Sentiment will be highlighted in the corresponding colour. Use the “<” and the “>” links to scroll through the Records.
If a Record was translated to English from another language you will see a toggle on the top right of the Record details. Flip this toggle to see the Record in the original language.
This process is the same for the Topics and Opinions views.