Generation of natural language notifications
Natural language notification generation techniques and system are described. In an implementation, natural language notifications are generated to provide insight into alerts related to a metric, underlying causes of the alert from other metrics, and relationships of the metric to other metrics. In this way, a user may gain this insight in an efficient, intuitive, and time effective manner.
1. In a digital medium environment to control output of a natural language notification regarding a service provider, a method implemented by at least one computing device, the method comprising:
obtaining, by the at least one computing device, metric data describing metrics involving interaction of a user population via a network with the service provider, the user population including multiple users;
generating, by the at least one computing device, statements describing a plurality of different data changes or trends of respective said metrics and relationships of the metrics to each other;
determining one or more user attributes of a user to which the natural language notification is to be output;
forming, by the at least one computing device, at least one natural language phrase based on the generated statements by rectifying a set of statement templates as including respective said statements, said rectifying including selecting, based on the one or more user attributes and from multiple different available levels of detail for describing the statements, a level of detail for describing the statements in the at least one natural language phrase; and
outputting, by the at least one computing device, the natural language notification that includes the at least one natural language phrase.
2. The method as described in claim 1 , wherein the metrics describe at least one metric-to-be-reported, in operation of the service provider or an insight into a change observed in the user interaction.
3. The method as described in claim 1 , wherein the generating of the statements is based at least in part on types of tags assigned to respective said metrics.
4. The method as described in claim 3 , wherein the types of tags include:
semantic tags describing a type of user interaction with the service provider;
functional tags that describe a type of information being tracked by a respective said metric;
customer journey tags describing a journey through digital content by a user in relation to a respective said metric; or
persona tags describing a type of user that is typically interested in a respective said metric.
5. The method as described in claim 1 , wherein the metric data describes metrics involving operation of the service provider as part of making a product or service available for conversion via the network.
6. The method as described in claim 1 , further comprising aggregating a plurality of said natural language phrases to form the natural language notification.
7. The method as described in claim 1 , further comprising referencing complementary information in the natural language notification based at least in part on the metrics included in the metric data.
8. The method as described in claim 1 , further comprising arranging text, order, layout, or dispatch of the natural language notification based on at least one user preference.
9. The method as described in claim 8 , wherein the at least one user preference is based on a user's role, profile, or usage information that is to receive the natural language notification.
10. The method as described in claim 1 , wherein the metric data describes the metrics through use of a value of the metric and a timestamp and the relationships of the metrics to each other include a cause behind at least one of the relationships based on the metric data over time.
11. In a digital medium environment to control output of a natural language notification regarding a service provider, a method implemented by at least one computing device, the method comprising:
obtaining, by the at least one computing device, metric data describing an alert involving a comparison of a metric of the service provider with a threshold;
generating, by the at least one computing device, statements describing data changes or trends of at least one other metric that has a likelihood of causing the alert of the metric;
forming, by the at least one computing device, at least one natural language phrase based on the generated statements by rectifying a set of statement templates after inclusion of respective said statements, said forming including:
determining that multiple natural language phrases are formable by said rectifying; and
selecting the at least one natural language phrase based on determining that the at least one natural language phrase provides more detail describing the generated statements than a different natural language phrase that is formable by said rectifying; and
outputting, by the at least one computing device, the natural language notification that includes the at least one natural language phrase.
12. The method as described in claim 11 , wherein the generating of the statements is based at least in part on types of tags assigned to respective said metrics.
13. The method as described in claim 12 , wherein the types of tags include:
semantic tags describing a type of user interaction with the service provider;
functional tags that describe a type of information being tracked by a respective said metric;
customer journey tags describing a journey through digital content by a user in relation to a respective said metric; or
persona tags describing a type of user that is typically interested in a respective said metric.
14. The method as described in claim 11 , wherein the generating of the statements is based at least in part on semantic analysis that identifies a relationship between the metric and the at least one other metric.
15. The method as described in claim 11 , wherein the generating of the statements is based at least in part on comparison of values of the at least one other metric to a threshold.
16. In a digital medium environment to control output of a natural language notification regarding a service provider, a system comprising:
a metric collection module implemented at least partially in hardware to obtain metric data describing metrics involving user interaction of a user population via a network with the service provider the user population including multiple users;
a statement generation module implemented at least partially in hardware to generate statements describing a plurality of different data changes or trends of respective said metrics and relationships of the metrics to each other;
a phrase generation module implemented at least partially in hardware to form at least one natural language phrase based on the generated statements by determining a set of fact values that are present in the statements, selecting a set of statement templates that apply to the fact values, and rectifying the set of statement templates after inclusion of respective said statements to form the at least one natural language phrase.
17. The system as described in claim 16 , wherein the metrics describe at least one anomaly in operation of the service provider.
18. The system as described in claim 16 , wherein the metric data describes metrics involving operation of the service provider as part of making a product or service available for conversion via the network.
19. The system as described in claim 16 , wherein the statement generation module is configured to generate the statements based at least in part on types of tags assigned to respective said metrics.
20. The system as described in claim 16 , wherein the metric data describes the metrics through use of a value of the metric and a timestamp and the relationships of the metrics to each other include a cause behind at least one of the relationships based on the metric data over time.