User-centered engagement analysis
View Patent ↗Techniques for analyzing user engagement are provided. The techniques can include obtaining event records for one or more user activities, aggregating the event records to a temporal resolution, accumulating computed counts for each quantized time, and computing percentiles for the accumulated counts for each quantized time. The aggregating can include quantizing time to the temporal resolution; and computing counts for the event records for each quantized time. The one or more activities can be defined for one or more behavior classes.
1. A method for analyzing user engagement comprising:
associating multiple user activities with a plurality of behavior classes, wherein a user activity is associated with one or more behavior classes;
obtaining event records for the multiple user activities;
quantizing time to a temporal resolution, resulting in a plurality of quantized times;
aggregating the event records, for each of the plurality of quantized times, for a behavior class of the plurality of behavior classes, wherein the aggregating comprises:
identifying, for the quantized time, event records for activity types associated with the behavior class, and
computing a count of the identified event records for activity types associated with the behavior class for the quantized time;
accumulating the computed counts of event records for activity types associated with the behavior class for each quantized time; and
computing percentiles for the accumulated counts of event records for activity types associated with the behavior class for each quantized time.
2. The method of claim 1 , wherein the plurality of behavior classes comprises consumption, creation, and interaction.
3. The method of claim 1 , wherein the event records comprise time, user id, user action, and artifact target.
4. The method of claim 2 , wherein:
the associating comprises associating viewed document and viewed blog post user activities with the consumption behavior class;
the behavior class of the plurality of behavior classes is the consumption behavior class; and
the aggregating comprises, for each of the plurality of quantized times, identifying event records for the viewed document and/or viewed blog post user activities.
5. The method of claim 2 , wherein:
the associating comprises associating create document and create blog post user activities with the creation behavior class;
the behavior class of the plurality of behavior classes is the creation behavior class; and
the aggregating comprises, for each of the plurality of quantized times, identifying event records for the create document and/or create blog post user activities associated with the creation behavior class.
6. The method of claim 1 , wherein the temporal resolution is one day.
7. The method of claim 1 , wherein the accumulating comprises:
summing the computed counts for each quantized time by user ID.
8. The method of claim 1 , further comprising:
charting the percentiles for the accumulated counts of event records for activity types associated with the behavior class for each quantized time.
9. The method of claim 8 , wherein the charting comprises charting 50, 70, and 90 percentiles for the accumulated counts over time.
10. A computing device adapted to perform a method for analyzing user engagement, the method comprising:
associating multiple user activities with a plurality of behavior classes, wherein a user activity is associated with one or more behavior classes;
obtaining event records for the multiple user activities;
quantizing time to a temporal resolution, resulting in a plurality of quantized times;
aggregating the event records, for each of the plurality of quantized times, for a behavior class of the plurality of behavior classes, wherein the aggregating comprises:
identifying, for the quantized time, event records for activity types associated with the behavior class, and
computing a counts of the identified event records for activity types associated with the behavior class for the quantized time;
accumulating the computed counts for each quantized time for the behavior class; and
computing percentiles for the accumulated counts for each quantized time, wherein the computing percentiles comprises:
stepping through each quantized time,
identifying, for each quantized time, accumulated computed counts for the behavior class, and
computing a percentile of the accumulated count at each quantized time for the behavior class, wherein the percentile is based on a user ID and accumulated counts.
11. The computing device of claim 10 , wherein the plurality of behavior classes comprises consumption, creation, and interaction.
12. The computing device of claim 10 , wherein the event records comprise time, user id, user action, and artifact target.
13. The computing device of claim 11 , wherein:
the associating comprises associating viewed document and viewed blog post user activities with the consumption behavior class; and
the aggregating comprises, for each of the plurality of quantized times, identifying event records for the viewed document and/or viewed blog post user activities associated with the consumption behavior class.
14. The computing device of claim 11 , wherein:
the associating comprises associating create document and create blog post user activities with the creation behavior class; and
the aggregating comprises, for each of the plurality of quantized times, identifying event records for the create document and/or create blog post user activities associated with the creation behavior class.
15. The computing device of claim 10 , wherein the temporal resolution is one day.
16. The computing device of claim 10 , wherein the accumulating comprises:
summing the computed counts for each quantized time by user ID.
17. The computing device of claim 10 , wherein the method further comprises:
charting the percentiles for the accumulated counts for each quantized time for the behavior class.
18. A computer-readable storage medium storing computer-executable instructions for causing a processor programmed thereby to perform a method comprising:
associating multiple user activities with a plurality of behavior classes, wherein a user activity is associated with one or more behavior classes;
obtaining event records for the multiple user activities;
quantizing time to a temporal resolution, resulting in a plurality of quantized times;
aggregating the event records, for each of the plurality of quantized times, for a behavior class of the plurality of behavior classes, wherein the aggregating comprises:
identifying, for the quantized time, event records for activity types associated with the behavior class, and
computing a counts of the identified event records for activity types associated with the behavior class for the quantized time;
accumulating the computed counts for each quantized time for the behavior class by user ID;
computing percentiles for the accumulated counts for each quantized time, wherein the computing percentiles comprises:
stepping through each quantized time,
identifying, for each quantized time, accumulated computed counts for the behavior class, and
computing a percentile of the accumulated counts at each quantized time for the behavior class, wherein the percentile is based on user ID and accumulated count; and
charting the percentiles for the accumulated counts for each quantized time for the behavior class.
19. The method of claim 7 , wherein the computing the percentile comprises:
defining the percentile as a number of summed counts at the quantized time.
20. The computing device of claim 7 , wherein the computing the percentile comprises:
defining a first percentile as a number of summed counts at the quantized time for a first user ID;
defining a second percentile as a number of summed events at the quantized time for a second user ID;
interpolating the percentile based on the first percentile and the second percentile.