Configuring Text Analytics for social media data

For companies that have purchased and implemented Medallia Social Feedback and Medallia Text Analytics, Medallia Experience Cloud can run Text Analytics on comments from review websites. Complete the steps in this topic to add social media data as a Text Analytics source.

Warning: As part of this procedure, you modify the K-field used as the Text Analytics MainScoreField to include the OverallScoreField used by Social Feedback. The MainScoreField must not be the same as the OverallScoreField configured for social media. If the fields are the same, completing this procedure can create a recursive loop for the calculations being performed by the K-fields, potentially making your Experience Cloud instance unusable.

Open the Reporting > Text Processing > Topic & Theme Tagging > Auto Tagging Pools screen in Setup.
In the MemberSurveyFields property, shuttle over the social media comment Q-field (q_socialmedia_comment).
On the Feedback > Social Media > Social Feedback screen, note the value of the OverallScoreField property.
On the Company > Profile > Company Settings screen, verify that the MainScoreField property has a different value than the OverallScoreField property you noted in the previous step.
If the values for the MainScoreField and OverallScoreField properties are different, proceed directly to the next step. If the values are the same, you must create a new K-field for the MainScoreField before proceeding to the next step.
Modify the K-field used for the MainScoreField to combine the main score from Experience Cloud survey feedback and the a_overall_score_with_social_media_5_buckets field, the key customer loyalty metric associated with social data.
For information about configuring the main score for legacy reports, see Configuring the impact score.
Note: Reports in Text Analytics reporting packages already account for data from Experience Cloud survey feedback and social feedback. For more information about these reports, Configuring the Generic Text Analytics reporting app.
Tip: Some alternative sets have a sequence number of -1 (so a score of 10 = 9). When modifying the K-field used for the MainScoreField property, first check the sequence number of the alternative set you are referencing to avoid a misrepresentation of scores.
Important: Changing the MainScoreField to include social data changes the comment count in any report that does not have the Feedback Type filter. Instead of including the number of comments from surveys only, the comment count increases to reflect the number of comments from surveys and social media. However, the comment count is unaffected in reports with the Feedback Type filter, since the user can choose to view comments from surveys or social media, but cannot view both types of feedback at the same time.
Example of a K-field:
```
function(){
	if (q_loyalty_recommend != null) {
		q_loyalty_recommend;
	}
	else if (a_overall_score_with_social_media_5_buckets != null) {
		seqnum(a_overall_score_with_social_media_5_buckets);
	}	
	else {
		null;
	}
			  }
```
Note: The social media score is based on a 5-point scale. In cases where the main score from surveys is based on a different scale, consider converting the social media score to a different scale. For example, the following excerpt from a K-field multiplies the value from a 5-point social media score field by 2:
```
...

	var socialScore = seqnum(a_overall_score_with_social_media_5_buckets);
	if (socialScore !== null) {
		return 2 * socialScore;
	}

...
```
Add the Feedback Type filter to your Text Analytics reports and dashboards, at the module level and dashboard level.
This filter allows users to specify that a report shows only data from a specific source, such as from social review websites.
For custom modules, add the FEEDBACK_TYPE plugin to the optionbox-filter element in the AA2 XML. For more information about working with custom modules and AA2 XML, see Custom Module.
Add social media reporting permissions to applicable roles.
For more information, see Reporting permissions.