Create/Edit a Dataset

Required security profile permissions: NICE inContact Interaction Analytics Edit

If you are an administrator and want to apply a custom category setThe structure of the transcript data your topic gleans, organized to aid in using the data for a specific purpose. Out-of-the-box category sets include Intent to Buy and Risk Aversion. to this new datasetA set of voice, email, or chat transcriptions from your contact center that spans a specific period of time and matches your specified filter criteria. that you have not already created, create the category set now.
If you are an administrator and want to use a custom workspace setA set of multiple workspaces grouped as a reusable template. When you first save a topic, you choose a dashboard set as the starting set of dashboards for the topic. with this dataset that you have not already created, create the workspace set now.
If you are creating a new dataset, click My DataAnalyze New Data. If you are modifying an existing dataset, click My Data, the name of the dataset you are editing, then Edit Topic.
Click the Channels drop-down and select all channels you want to include in the dataset. The preceding image selects multiple channels.
Give the dataset a descriptive Dataset Name. Specify the Period you want the dataset to include and apply the Category Set and Workspace Set you want.

If you want to filter the dataset data to only include contacts that include or exclude certain keywordsAn individual word you enter into NICE inContact Interaction Analytics Pros to flag it as a word with special significance or as a word to use as criteria for including or excluding data from a topic or widget., phrasesA combination of words, contained in double quotes (""), you enter into NICE inContact Interaction Analytics Pro to flag it as a phrase with special significance or as a phrase to use as criteria for including or excluding data from a topic or widget., entitiesA term that represents a theme or concept. NICE inContact Interaction Analytics Pro intelligently searches for related keywords and context to gather instances of the concept rather than searching for every specific instance of a keyword. For example, the 'sound quality' entity applies to transcripts where people say things like "I can't hear you" or "you're breaking up". You can create custom entities specific to your organization in the company profile., sentimentsThe overall mood of the contact, or the result of the call as determined by analysis of words, phrases, and context of the transcript. An interaction can be positive (blue), negative (red), frustrated, mixed (dark gray), or neutral (light gray)., metrics, or a combination of these things, use the All of These, At Least One of These, and None of These fields as needed:

Click and drag sentiment, metric, frustrated, and unresolved tags from Add criteria to the preferred fields and configure them as needed.

See more information about configuring the Add criteria tags

Tag	Description
Sentiment	Pulls data into the dataset based on the sentiment assigned to the interactions. After you drag a sentiment tag to the preferred field, click the agent/contact icon on the left side of the tag (indicated by a headset or person icon, or both) to toggle the options between agent sentiment, contact sentiment, or agent and contact sentiment. Click the down arrow on the right side of the tag to open a configuration pop-up. Select the Type of sentiment you want to use as criteria for your dataset between End Sentiment at the conclusion of the interaction or the Overall sentiment throughout the interaction. Select the preferred Sentiment from Positive, Negative, Neutral, and Mixed. Click Done.
Metric	Pulls data into the dataset based on whether it meets a certain metric value or threshold. After you drag a metric tag to the preferred field, configure the metric criteria in the pop-up. If a pop-up did not appear when you dropped the tag or you closed it, click the down arrow in the tag to open it again. Set the Field to the metric you want use as the criteria. See the Metric Widget configuration dataset for details about each available metric. If you selected a metric that produces a numeric value, select an operator from the drop-down under Value from Equals, Greater than, Less than, Greater than or equal to, or Less than or equal to and type the number threshold you want the operator to compare with the metric in the text field. If you selected a metric that produces a string value, type the value you want the metric to match in the Value text box. If you selected a metric that produces a boolean value, select True or False from the drop-down. Click Done.
Frustrated	Pulls data into the dataset based on whether the contact became frustrated during the interaction.
Unresolved	Pulls data into the dataset based on the resolution status (unresolved or resolved) at the end of a contact. After you drag an unresolved tag to the preferred field, click the down arrow on the right side of the tag. Specify in the Type field whether you want Resolved or Unresolved contacts to be the criteria. Click Done.

Enter text directly into the Find contacts with fields to enter keywords, phrases, or entities.

For each tag you type, you can specify the following:
- Whether the keyword, phrase, or entity should appear in the agent or contact side of the interaction, or both, by clicking the agent or contact icon on the left side of the tag (indicated by a headset or person icon) until the icon corresponding to the criteria you want appears.
- Whether the keyword, phrase, or entity should be an exact match to what you entered or a variation of what you entered, by clicking the down arrow on the right. An exact match only returns the term or phrase exactly as you entered it. If you choose to match with variations it would also return variations of that word, for example if you enter activate, it could return interactions that include activate, activates, activating, activated, and so on.
When you enter a phrase, clicking the down arrow also allows you to specify how close the words in the phrase must be to each other for the contact to be considered a match on the phrase. The default is Exact Phrase, meaning the words must be together and in the right order. You can also specify that the words must be one or fewer words apart to eight or fewer words apart and not necessarily consecutive. For example, if you enter the phrase "rude agent" and select a proximity of 3 Words, a transcript that says "an agent was disrespectful and rude" would count.
Click Save.