Understanding the AWS Glue data catalog tables

The following tables list and describe the columns, data types, and elements in an Amazon Chime SDK call analytics Glue data catalog.

call_analytics_metadata

Column name	Data type	Elements	Definition
time	string		Event generation timestamp ISO 8601.
detail-type	string		Feature type related to service-type.
service-type	string		Name of the AWS service, VoiceAnalytics or CallAnalytics.
detail-subtype	string		Used for Recording and CallAnalyticsMetadata detail-types.
callevent-type	string		Event type associated with SIP, such as Update, Pause, Resume
mediaInsightsPipelineId	string		Amazon Chime SDK media insights pipeline ID.
metadata	string	voiceConnectorId	The Amazon Chime SDK Voice Connector ID.
		callId	The call ID of the participant for the associated usage.
		transactionId	The transaction ID of the call.
		fromNumber	E.164 origination phone number.
		toNumber	E.164 destination phone number.
		direction	Direction of the call, Outbound or Inbound.
		oneTimeMetadata.s3RecordingUrl	Amazon S3 bucket URL of the media object emitted by Transcribe Call Analytics.
		oneTimeMetadata.s3RecordingUrlRedacted	Amazon S3 bucket URL of the redacted media object emitted by Transcribe Call Analytics.
		oneTimeMetadata.siprecMetadata	SIPREC Metadata in XML format associated with the call.
		oneTimeMetadata.siprecMetadataJson	SIPREC Metadata in JSON format associated with the call.
		oneTimeMetadata.InviteHeaders	Invite headers.

call_analytics_recording_metadata

Column name	Data type	Elements	Definition
time	string		Event generation timestamp ISO 8601.
detail-type	string		Feature type related to service-type.
service-type	string		Name of the AWS service, VoiceAnalytics or CallAnalytics.
detail-subtype	string		Used for Recording and CallAnalyticsMetadata detail-types.
callevent-type	string		Event type associated with SIP
mediaInsightsPipelineId	string		Amazon Chime SDK media insight pipeline ID.
s3MediaObjectConsoleUrl	string		S3 Bucket URL of the media object.
metadata	string	voiceConnectorId	The Amazon Chime SDK Voice Connector ID.
		callId	The call ID of the participant for the associated usage.
		transactionId	The transaction ID of the call.
		fromNumber	E.164 origination phone number.
		toNumber	E.164 destination phone number.
		direction	Direction of the call, Outbound or Inbound.
		voice enhancement	Feature subtype related to service-type.
		oneTimeMetadata.siprecMetadata	SIPREC Metadata in XML format associated with the call.
		oneTimeMetadata.siprecMetadataJson	SIPREC Metadata in JSON format associated with the call.
		oneTimeMetadata.InviteHeaders	Invite headers.

transcribe_call_analytics

Column name	Data type	Elements	Definition
time	string		Event generation timestamp ISO 8601.
detail-type	string		Feature type related to service-type.
service-type	string		Name of the AWS service, VoiceAnalytics or CallAnalytics.
mediaInsightsPipelineId	string		Amazon Chime SDK media insight pipeline ID.
metadata	string	voiceConnectorId	The Amazon Chime Voice Connector ID.
		callId	The call ID of the participant for the associated usage.
		transactionId	The transaction ID of the call.
		fromNumber	E.164 origination phone number.
		toNumber	E.164 destination phone number.
		direction	Direction of the call, Outbound or Inbound.
UtteranceEvent	struct	UtteranceId	The unique identifier associated with the specified UtteranceEvent.
		IsPartial	Indicates whether the segment in the UtteranceEvent is complete (FALSE) or partial (TRUE).
		ParticipantRole	Provides the role of the speaker for each audio channel, either CUSTOMER or AGENT.
		BeginOffsetMillis	The time, in milliseconds, from the beginning of the audio stream to the start of the UtteranceEvent.
		EndOffsetMillis	The time, in milliseconds, from the beginning of the audio stream to the start of the UtteranceEvent.
		Transcript	Contains transcribed text.
		Sentiment	Provides the sentiment detected in the specified segment.
		Items.beginoffsetmillis	The start time, in milliseconds, of the transcribed item.
		Items.endoffsetmillis	The end time, in milliseconds, of the transcribed item.
		Items.itemtype	The type of item identified. Options: PRONUNCIATION (spoken words) and PUNCTUATION.
		Items.content	The word or punctuation that was transcribed.
		Items.confidence	The confidence score associated with a word or phrase in your transcript. Scores are values between 0 and 1. A larger value indicates a higher probability that the identified item correctly matches the item spoken in your media.
		Items.vocabularyfiltermatch	Indicates whether the specified item matches a word in the vocabulary filter included in your request. If true, there is a vocabulary filter match.
		Items.stable	The partial result stabilization is enabled, Stable indicates whether the specified item is stable (true) or if it may change when the segment is complete (false).
		IssuesDetected.characteroffsets_begin	Provides the character count of the first character where a match is identified. For example, the first character associated with an issue or a category match in a segment transcript.
		IssuesDetected.characteroffsets_end	Provides the character count of the last character where a match is identified. For example, the last character associated with an issue or a category match in a segment transcript.
		Entities.beginoffsetmillis	The start time, in milliseconds, of the utterance that was identified as PII.
		Entities.endoffsetmillis	The end time, in milliseconds, of the utterance that was identified as PII.
		Entities.category	The category of information identified. The only category is PII.
		Entities.type	The type of PII identified. For example, NAME or CREDIT_DEBIT_NUMBER.
		Entities.content	The word or words identified as PII.
		Entities.confidence	The confidence score associated with the identified PII entity in your audio. Confidence scores range between 0 and 1. A larger value indicates a higher probability that the identified entity correctly matches the entity spoken in your media.

transcribe_call_analytics_category_events

Column name	Data type	Elements	Definition
time	string		Event generation timestamp ISO 8601.
detail-type	string		Feature type related to service-type.
service-type	string		Name of the AWS service, VoiceAnalytics or CallAnalytics.
mediaInsightsPipelineId	string		Amazon Chime SDK media insight pipeline ID.
metadata	string	voiceConnectorId	The Amazon Chime Voice Connector ID.
		callId	The call ID of the participant for the associated usage.
		transactionId	The transaction ID of the call.
		fromNumber	E.164 origination phone number.
		toNumber	E.164 destination phone number.
		direction	Direction of the call, Outbound or Inbound.
CategoryEvent	array	MatchedCategories	Lists the matches in the categories defined by the user.

transcribe_call_analytics_post_call

Column name	Data type	Elements	Definition
JobStatus	string		Event generation timestamp ISO 8601.
LanguageCode	string		Feature type related to service-type.
Transcript	struct	LoudnessScores	Measures the volume at which each participant is speaking. Use this metric to see if the caller or the agent is speaking loudly or yelling, which often indicates anger. This metric is represented as a normalized value (speech level per second of speech in a given segment) on a scale from 0 to 100, where a higher value indicates a louder voice.
		Content	Contains transcribed text.
		Id	The unique identifier associated with the specified UtteranceEvent.
		BeginOffsetMillis	The time, in milliseconds, from the beginning of the audio stream to the start of the UtteranceEvent.
		EndOffsetMillis	The time, in milliseconds, from the beginning of the audio stream to the start of the UtteranceEvent.
		Sentiment	Provides the sentiment detected in the specified transcript segment.
		ParticipantRole	Provides the role of the speaker for each audio channel, either CUSTOMER or AGENT.
		IssuesDetected.CharacterOffsets.Begin	Provides the character offset to the first character where a match is identified. For example, the first character associated with an issue in a transcript segment.
		IssuesDetected.CharacterOffsets.End	Provides the character offset to the last character where a match is identified. For example, the last character associated with an issue in a transcript segment.
		OutcomesDetected.CharacterOffsets.Begin	Provides the outcome, or resolution, identified in the call.
		OutcomesDetected.CharacterOffsets.End
		ActionItemsDetected.CharacterOffsets.Begin	Lists any action items identified in the call.
		ActionItemsDetected.CharacterOffsets.End
AccountId	string		The AWS account Id
Categories	struct	MatchedCategories	Lists the matched categories.
		MatchedDetails	Lists the time, in milliseconds, from the beginning of the audio stream to when the Match in the category was detected.
Channel	string	Channel	Indicates a Voice channel.
Participants	array	ParticipantRole	Provides the role of the speaker for each audio channel, CUSTOMER or AGENT.
ConversationCharacteristics	struct	NonTalkTime	Measures periods of time that do not contain speech. Use this metric to find long periods of silence, such as a customer on hold for an excessive amount of time.
		Interruptions	Measures if and when one participant cuts off the other participant mid-sentence. Frequent interruptions may be associated with rudeness or anger, and could correlate to negative sentiment for one or both participants.
		TotalConversationDurationMillis	Total length of the conversation.
		Sentiment.OverallSentiment.AGENT	OverallSentiment label for the Agent.
		Sentiment.OverallSentiment.CUSTOMER	OverallSentiment label for the Customer.
		Sentiment.SentimentByPeriod.QUARTER.AGENT	Sentiment labels for each quarter for the Agent.
		Sentiment.SentimentByPeriod.QUARTER.CUSTOMER	Sentiment labels for each quarter for the Customer.
		TalkSpeed	Measures the speed at which both participants are speaking. Comprehension can be affected if one participant speaks too quickly. This metric is measured in words per minute.
		TalkTime	Measures the amount of time (in milliseconds) each participant spoke during the call. Use this metric to help identify if one participant is dominating the call or if the dialogue is balanced.
SessionId	string		SessionId for the call
ContentMetadata	string		Field that labels raw vs. redacted content per the customer specified configuration.

transcribe

Column name	Data type	Elements	Definition
time	string		Event generation timestamp ISO 8601.
detail-type	string		Feature type related to service-type.
service-type	string		Name of the AWS service, VoiceAnalytics or CallAnalytics.
mediaInsightsPipelineId	string		Amazon Chime SDK media insight pipeline ID.
metadata	string	voiceConnectorId	The Amazon Chime Voice Connector ID.
		callId	The call ID of the participant for the associated usage.
		transactionId	The transaction ID of the call.
		fromNumber	E.164 origination phone number.
		toNumber	E.164 destination phone number.
		direction	Direction of the call, Outbound or Inbound.
TranscriptEvent	struct	ResultId	The unique identifier for the Result.
		StartTime	The start time, in milliseconds, of the Result.
		EndTime	The end time, in milliseconds, of the Result.
		IsPartial	Indicates whether the segment is complete. If IsPartial is true, the segment is not complete. Otherwise, the segment is complete.
		ChannelId	The ID of the channel associated with the audio stream.
		Alternatives.Entities	Contains entities identified as personally identifiable information (PII) in your transcription output.
		Alternatives.Items.Confidence	The confidence score associated with a word or phrase in your transcript. Confidence scores are values between 0 and 1. A larger value indicates a higher probability that the identified item correctly matches the item spoken in your media.
		Alternatives.Items.Content	The transcribed word or punctuation mark.
		Alternatives.Items.EndTime	The end time, in milliseconds, of the transcribed item.
		Alternatives.Items.Speaker	If speaker partitioning is enabled, Speaker labels the speaker of the specified item.
		Alternatives.Items.Stable	If partial result stabilization is enabled. Stable indicates whether the specified item is stable (true) or if it may change when the segment is complete (false).
		Alternatives.Items.StartTime	The start time, in milliseconds, of the transcribed item.
		Alternatives.Items.Type	The type of item identified. Options: PRONUNCIATION (spoken words) and PUNCTUATION.
		Alternatives.Items.VocabularyFilterMatch	Indicates whether the specified item matches a word in the vocabulary filter included in your request. If true, there is a vocabulary filter match.
		Alternatives.Transcript	Contains transcribed text.

voice_analytics_status

Column name	Data type	Elements	Definition
time	string		Event generation timestamp ISO 8601.
detail-type	string		Feature type related to service-type.
service-type	string		Name of the AWS service, VoiceAnalytics or CallAnalytics.
source	string		AWS service that produces the event.
account	string		AWS Account ID.
region	string		AWS Account Region.
version	string		Version of the event schema.
id	string		Unique ID of the event
detail	struct	taskId	Unique ID of the task.
		isCaller	Indicates whether the participant is caller or not.
		streamStartTime	Start time of the stream.
		transactionId	The transaction ID of the call.
		voiceConnectorId	The Amazon Chime Voice Connector ID.
		callId	The call ID of the participant for the associated usage.
		detailStatus	Detailed feature type related to service-type.
		statusMessage	Status of task ID success or failure.
		mediaInsightsPipelineId	Amazon Chime SDK media insight pipeline ID. This field is populated only for speaker search tasks started through the Media Pipelines SDK, not the Voice SDK.
		sourceArn	The resource ARN for which the task is run on
		streamArn	The Kinesis Video Stream ARN that the task is run for. This field is populated only for speaker search tasks started through the Media Pipelines SDK, not the Voice SDK.
		channelId	The channel of the streamArn that the task is run for. This field is populated only for speaker search tasks started through the Media Pipelines SDK, not the Voice SDK.
		speakerSearchDetails.voiceProfileId	ID of a voice profile enrolled whose voice embedding matches closely with the speaker in the call.
		speakerSearchDetails.confidenceScore	Number between [0, 1] where a larger number means the machine learning model is more confident about the voice profile match.

speaker_search_status

Column name	Data type	Elements	Definition
time	string		Event generation timestamp ISO 8601.
detail-type	string		Feature type related to service-type.
service-type	string		Name of the AWS service, VoiceAnalytics or CallAnalytics.
source	string		AWS service that produces the event.
account	string		AWS Account ID.
region	string		AWS Account Region.
version	string		Version of the event schema.
id	string		Unique ID of the event
detail	struct	taskId	Unique ID of the task.
		isCaller	Indicates whether the participant is caller or not.
		transactionId	The transaction ID of the call. This field is populated if the task originates from a call made through a Voice Connector.
		voiceConnectorId	The Amazon Chime Voice Connector ID. This field is populated if the task originates from a call made through a Voice Connector.
		mediaInsightsPipelineId	The media insights pipeline ID. This field is populated only for speaker search tasks started through the Media Pipelines SDK, not the Voice SDK.
		sourceArn	The resource ARN for which the task is run on.
		streamArn	The Kinesis Video Stream ARN that the task is run for. This field is populated only for speaker search tasks started through the Media Pipelines SDK, not the Voice SDK.
		channelId	The channel of the streamArn that the task is run for. This field is populated only for speaker search tasks started through the Media Pipelines SDK, not the Voice SDK.
		participantRole	The participant role associated with the channelId in the streamArn. This field is populated only for speaker search tasks started through the Media Pipelines SDK, not the Voice SDK.
		detailStatus	Detailed feature type related to service-type.
		statusMessage	Status of task ID, success or failure.
		speakerSearchDetails.voiceProfileId	ID of a voice profile enrolled whose voice embedding matches closely with the speaker in the call.
		speakerSearchDetails.confidenceScore	Number between [0, 1] where a larger number means the machine learning model is more confident about the voice profile match.

voice_tone_analysis_status

Column name	Data type	Elements	Definition
time	string		Event generation timestamp ISO 8601.
detail-type	string		Feature type related to service-type.
service-type	string		Name of the AWS service, VoiceAnalytics or CallAnalytics.
source	string		AWS service that produces the event.
account	string		AWS Account ID.
region	string		AWS Account Region.
version	string		Version of the event schema.
id	string		Unique ID of the event
detail	struct	taskId	Unique ID of the task.
		isCaller	Indicates whether the participant is caller or not.
		transactionId	The transaction ID of the call. This field is populated if the task originates from a call made through a Voice Connector.
		voiceConnectorId	The Amazon Chime Voice Connector ID. This field is populated if the task originates from a call made through a Voice Connector.
		mediaInsightsPipelineId	The media insights pipeline ID. This field is populated only for speaker search tasks started through the Media Pipelines SDK, not the Voice SDK.
		sourceArn	The resource ARN for which the task is run on.
		streamArn	The Kinesis Video Stream ARN that the task is run for. This field is populated only for speaker search tasks started through the Media Pipelines SDK, not the Voice SDK.
		channelId	The channel of the streamArn that the task is run for. This field is populated only for speaker search tasks started through the Media Pipelines SDK, not the Voice SDK.
		participantRole	The participant role associated with the channelId in the streamArn. This field is populated only for speaker search tasks started through the Media Pipelines SDK, not the Voice SDK.
		statusMessage	Status of task ID success or failure.
		voiceToneAnalysisDetails.startFragmentNumber	Starting fragment number associated with the streamArn.
		voiceToneAnalysisDetails.currentAverageVoiceTone.startTime	Starting timestamp in ISO8601 format for the speaker's call audio that the current average sentiment is based on.
		voiceToneAnalysisDetails.currentAverageVoiceTone.endTime	Ending timestamp in ISO8601 format for the speaker's call audio that the current average sentiment is based on.
		voiceToneAnalysisDetails.currentAverageVoiceTone.beginOffsetMillis	Beginning offset in milliseconds from the starting fragment for the speaker's call audio that the current average sentiment is based on.
		voiceToneAnalysisDetails.currentAverageVoiceTone.endOffsetMillis	Ending offset in milliseconds from the starting fragment for the speaker's call audio that the current average sentiment is based on.
		voiceToneAnalysisDetails.currentAverageVoiceTone.voiceToneScore.positive	Probabilistic likelihood between [0, 1] that the speaker's sentiment is positive.
		voiceToneAnalysisDetails.currentAverageVoiceTone.voiceToneScore.negative	Probabilistic likelihood between [0, 1] that the speaker's sentiment is negative.
		voiceToneAnalysisDetails.currentAverageVoiceTone.voiceToneScore.neutral	Probabilistic likelihood between [0, 1] that the speaker's sentiment is neutral.
		voiceToneAnalysisDetails.currentAverageVoiceTone.voiceToneLabel	Label with highest probability for the average voice tone score.
		voiceToneAnalysisDetails.overallAverageVoiceTone.startTime	Starting timestamp in ISO8601 format for the speaker's call audio that the overall average sentiment is based on.
		voiceToneAnalysisDetails.overallAverageVoiceTone.endTime	Ending timestamp in ISO8601 format for the speaker's call audio that the overall average sentiment is based on.
		voiceToneAnalysisDetails.overallAverageVoiceTone.beginOffsetMillis	Beginning offset in milliseconds from the starting fragment for the speaker's call audio that the overall average sentiment is based on.
		voiceToneAnalysisDetails.overallAverageVoiceTone.endOffsetMillis	Ending offset in milliseconds from the starting fragment for the speaker's call audio that the overall average sentiment is based on.
		voiceToneAnalysisDetails.overallAverageVoiceTone.voiceToneScore.positive	Probabilistic likelihood between [0, 1] that the speaker's sentiment is positive.
		voiceToneAnalysisDetails.overallAverageVoiceTone.voiceToneScore.negative	Probabilistic likelihood between [0, 1] that the speaker's sentiment is negative.
		voiceToneAnalysisDetails.overallAverageVoiceTone.voiceToneScore.neutral	Probabilistic likelihood between [0, 1] that the speaker's sentiment is neutral.
		voiceToneAnalysisDetails.overallAverageVoiceTone.voiceToneLabel	Sentiment label (positive, negative, or neutral) with the highest sentiment score.