Create a Text Analytics Classic Dashboard

Customize and discover more about your Open Ends responses with configurable metrics and filters.

1. Open the Analytics app.
2. Create a new dashboard:
   • On the Home page, click Create a Dashboard.
   • In the Analytics navigation menu, click Dashboards and then click New Dashboard.
3. Enter a name for your dashboard and click Next.
4. In Set Dashboard Access Controls, specify which users can access the dashboard.

   The access settings you can apply depend on the user role you have been assigned.

   Access Setting	User Roles	Description
   Everyone	Power User, Author	All other users in the application instance can access the dashboard.
   Admin Only	Admin	Only Admin users can access the dashboard.
   Teams	Admin, Power User, Author	All users that belong to the assigned team(s) can access the dashboard. If you belong to multiple teams, you can assign the dashboard to all of the teams you belong to, or just one or more specific teams.

   Note:

   • The access settings options are displayed if your account has the Can change access settings permission applied.
   • If your account does not have this permission, the dashboard will be assigned to the teams you belong to, or if you don't belong to any teams it will be available to everyone.
   • Admins have access to all dashboards, regardless of access settings.
5. Click Next.
6. Click Classic Dashboards and click Next.
7. In the Select a Template page, select Text Analytics and click Next.
8. Select a dataset from the list and then click Next.
9. Find the results of the data reflected in the following tiles templates:
   Note: Tile templates have the flexibility to be edited to better suit the data.

   Tiles created from Open Ends:

   Tile		Description
   Taxonomy		The taxonomy tile displays a list of categories Open End responses fall under. Categories are created hierarchically with multiple levels of subcategories where Open End responses are grouped together and categorized. Note: The Taxonomy tile requires the community to have a taxonomy model set up. Please contact Professional Service for more information.
   Sentiment Over Time		Sentiments are shown at the response level over a period of six months. Filter by sentiments to see how the sentiments shift month by month.
   Total Number of Text Responses		The total number of text responses from all Open End questions. Note: The number may look inflated compared to the number of responses in NPS®1 By Category because it contains the total number of text responses from Open End questions as well.
   Sentiment by Category		Sentiments are extracted from each Open End response at the sentence level. For example, if a response is "The product is great and easy to use.", it is categorized as positive.
   Key Phrase Sentiment Analysis		The full response is displayed along with key phrases sorted into either positive, negative, or both sentiments. The key phrases selected best describe the subject of the Open End response and multiple key phrases can be extracted from each response. For example, with a sentence such as, "The product is great! But the price is too expensive.", words like "product" will be categorized as positive and "price" as negative.
   Key Phrases Word Cloud		A word cloud of key phrases where the size is proportional to the frequency of the key phrase. Hover over a key phrase to see how many times it has been used overall. A key phrase is counted every time it is mentioned. For example, if the key phrase "issue" is mentioned three times in one sentence it will count three times.
   Sentiment By Key Phrase		The top ten most frequent key phrases found in Open Ends are broken down by sentiments. Sentiments are split into three categories, positive, neutral, and negative. Hover over each sentiment to see how many times a key phrase is mentioned. Note: Recall, a measure of the completeness of the analysis, is less than 100%. This means that not all key phrases can be clearly associated with a sentiment. For example, the algorithm may not assign the phrase "riot police" a negative sentiment, even though a human likely would. Only key phrases successfully associated with a sentiment will be displayed.

10. Use the existing filters for a refined view of the dashboard.
    
11. Optional: Create cross-filters to see how the data affects other variables.
12. Add additional visualizations to the dashboard.

    The following fields grouped under Text Analytics in the Dimensions list are specific to text analytics dashboards and can be added to your visualizations.

    

    Field	Description
    Category - name	The actual category and subcategories predicted. (Category/Sub-category1/Sub-category2/...).
    Category - score	The relevance score of the predicted category.
    Full text	The full text of the dataset open-end response field used in the analysis. Includes one record for each question response.
    Full text - Language	The language used in the survey.
    Full text - Sentiment Label	The label of the sentiment model. Positive, negative, or neutral.
    Full text - Sentiment Score	The score of the sentiment model. Sentiment score from -1 (completely negative) to 0 (neutral) to 1 (completely positive).
    Key phrase - count	The count of the occurrences of the detected keyword.
    Key phrase - relevance	The relevance score of the detected keyword. Duplicates are not counted, so if someone says "I will check if the check was cashed" it will count "check" only once.
    Key phrase - sentiment label	The label of the sentiment of the detected keyword. Positive, negative, or neutral.
    Key phrase - sentiment score	The sentiment score of the detected keyword. Sentiment score from -1 (completely negative) to 0 (neutral) to 1 (completely positive).
    Key phrase - text	The text of the detected keyword. Contains all key phrases identified.
    Text Fields - Field Name	The dataset field name associated with the full text or keyphrase.

    The following fields grouped under Text Analytics in the Measures list are specific to text analytics dashboards and can be added to your visualizations.

    

    Measure	Description
    Full text - Response Count	Counts each full text entry. This differs from the standard Count field which counts each survey response. For example, if you have a survey with 3 open-ends and 10 people complete it, the Full text - Response Count will be 30 and Count will be 10.
    Full text - Response Percent	The response count for a full text entry divided by the total response count. For example, 10% of people answered "no comment" to an open-ended question.
    Key phrase - count with duplicates	Unlike Key phrase - count this includes duplicates. If the key phrase is "I will check if the check was cashed", it will count "check" twice.
    Key phrase - text concat	Concatenates all key phrases into one long sentence separated by commas.

13. Download the dashboard.
14. Download a tile.