Systems and methods for cleansing automated robotic traffic from sets of usage logs

1. A method of cleansing data generated by one or more servers in response to interactions with users, the method comprising:

retrieving, via one or more servers, metadata in response to a first set of requests corresponding to a first event type;

receiving, via the one or more servers, a second set of requests in response to actuation of a link embedded in the metadata, at least one full record associated with the metadata being retrieved by the one or more servers from a source repository in response to one of the second set of requests, the second set of requests corresponding to a plurality of event types excluding the first event type;

logging, by the one or more servers, the first and second sets of requests as events;

determining, by the one or more servers, a relationship between the events based on the first and plurality of event types;

classifying, by the one or more servers, the first and second sets of requests based on the relationship as either corresponding to human behavior or automated software robot behavior;

in response to classifying the requests as corresponding to human behavior, modifying a metric associated with the at least one full record by the one or more servers;

in response to classifying the requests as corresponding to software robot behavior, leaving the metric unchanged by the one or more servers; and

adjusting subsequent discovery of the metadata associated with the at least one full record by the one or more servers based on the metric.

2. The method of claim 1 , wherein the first event type corresponds to investment events and at least one of the plurality of event types corresponds to payoff events, and

wherein the relationship comprises determining an arrival rate of the payoff events relative to the investment events.

3. The method of claim 2 , wherein classifying the first and second sets of requests comprises determining whether the arrival rate of the payoff events exceeds an arrival threshold.

4. The method of claim 1 , wherein classifying the first and second sets of requests comprises determining whether a time between the second set of requests is substantially identical.

5. The method of claim 1 , further comprising determining an intentionality associated with the events.

6. The method of claim 5 , wherein determining the intentionality comprises:

determining from the events a quantity of the requests in the first set of requests that were submitted which did not result in receipt of at least one of the requests in the second set of requests.

7. The method of claim 6 , wherein determining the intentionality further comprises:

determining a natural log of a quantity of the requests in the first set of requests that were submitted which did not result in receipt of at least one of the requests in the second set of requests; and

multiplying the natural log of the quantity by a multiplication factor.

8. The method of claim 7 , wherein the multiplication factor comprises a constant value divided by an arrival rate of the second set of requests.

9. The method of claim 7 , wherein determining the intentionality comprises:

dividing a product of the natural log of the quantity and the multiplication factor by a mathematical representation of an attention span of a human user.

10. The method of claim 1 , wherein the first and second sets of requests are classified as corresponding to human behavior and the method further comprises:

generating a work-level usage metric based on the second set of requests.

11. The method of claim 10 , further comprising generating forecast data based on the work-level usage metric.

12. The method of claim 10 , further comprising generating trend data based on the work-level usage metric.

13. The method of claim 10 , further comprising generating recommendation data based on the work-level usage metric.

14. The method of claim 10 , further comprising generating rank data based on the work-level usage metric.

15. A system for cleansing data generated by one or more servers in response to interactions with users, the system comprising:

a computer storage device storing

one or more servers programmed to:

retrieve metadata in response to a first set of requests corresponding to a first event type;

receive a second set of requests in response to actuation of a link embedded in the metadata, at least one full record associated with the metadata being retrieved by the one or more servers from a source repository in response to one of the second set of requests, the second set of requests corresponding to a plurality of event types excluding the first event type;

log the first and second sets of requests as events;

determine a relationship between the events based on the first and plurality of event types;

classify the requests based on the relationship as either corresponding to human behavior or automated software robot behavior;

modify a metric associated with the at least one full record in response to classification of the requests as corresponding to human behavior;

leave the metric unchanged in response to classifying the requests as corresponding to software robot behavior; and

adjust subsequent discovery of the metadata associated with the at least one full record based on the metric.

16. The system of claim 15 , wherein the first event type corresponds to investment events and at least one of the plurality of event types corresponds to payoff events,

wherein the relationship comprises determining an arrival rate of the payoff events relative to the investment events, and

wherein classifying the first and second sets of requests comprises determining whether the arrival rate of the payoff events exceeds an arrival threshold.

17. The system of claim 1 , wherein the one or more servers are programmed to classify the first and second sets of requests comprises determining whether a time between the second set of requests is substantially identical.

18. The system of claim 1 , wherein the one or more servers are programmed to determine the relationship by determining an intentionality associated the events based on a quantity of the requests in the first set of requests that were submitted which did not result in receipt of at least one of the requests in the second set of requests, a natural log of a quantity of the requests in the first set of requests that were submitted which did not result in receipt of at least one of the requests in the second set of requests, and multiplying the natural log of the quantity by a multiplication factor.

19. The system of claim 18 , wherein the multiplication factor comprises a constant value divided by an arrival rate of payoff events in the usage log.

20. The system of claim 15 , wherein the first and second sets of requests are classified as corresponding to human behavior and the one or more servers are programmed to generate a work-level usage metric based on the types of events, and

wherein the one or more servers are programmed to generate at least one of forecast data or trend data based on the work-level usage metric.

21. A non-transitory computer-readable medium comprising instructions, wherein execution of the instructions by one or more processing devices causes the one or more processing devices to:

retrieve metadata in response to a first set of requests corresponding to a first event type;

receive a second set of requests in response to actuation of a link embedded in the metadata, at least one full record associated with the metadata being retrieved by the one or more servers from a source repository in response to one of the second set of requests, the second set of requests corresponding to a plurality of event types excluding the first event type;

log the first and second sets of requests as events;

determine a relationship between the events based on the first and plurality of event types;

classify the requests based on the relationship as either corresponding to human behavior or automated software robot behavior;

modify a metric associated with the at least one full record in response to classification of the requests as corresponding to human behavior;

leave the metric unchanged in response to classifying the requests as corresponding to software robot behavior; and

adjust subsequent discovery of the metadata associated with the at least one full record based on the metric.