Real time audience forecasting
View Patent ↗A system, method, apparatus and processor readable media are described for real-time prediction of an advertising audience volume through analysis of historical audience data, and tuning of the predicted audience volume. Embodiments enable a user to specify a query for audience volume prediction. Such a query may be a Boolean combination of various audience categories. A time range may be determined that indicates the amount of historical data that is to be analyzed to make the audience volume prediction in real time. Employing the user-specified query, an audience volume prediction may be provided for a future time period, based on an analysis of retrieved historical audience data for the time range. Embodiments may also enable a user to tune the predicted audience volume through modification of the query through one or more iterations.
1. A computer implemented method for generating information regarding an advertising audience volume, comprising using a processor to perform actions over a network, the actions comprising:
processing a request for a real time prediction that forecasts the advertising audience volume over an editable future time period, the request including at least a query, wherein the request identifies the editable future time period and wherein the request corresponds to a specified level of service and quality of service that determines a number of prediction servers for processing the request;
managing, at a prediction server, a speed of processing the request at least by storing a portion of historical data in memory and another portion of the historical data in one or more persistent storage devices based in part or in whole upon frequencies of accessing the historical data;
evaluating the query on the historical data over the one or more past time periods, by:
identifying historical advertising audience volumes that would have been reached had the query been executed during the one or more past time periods;
retrieving a set of historical data by sampling the historical advertising audience volumes for reducing an amount of the historical data for processing the request, wherein a level of service and a quality of service are balanced against one another by changing at least a sampling rate for sampling the historical advertising audience volumes and a number of employed servers to satisfy the user specified level of service and quality of service; and
performing smoothing on the set of historical data retrieved from the historical advertising audience volumes based at least in part upon one or more data sources for the set of historical data and temporal differences in the one or more past time periods; and
generating the real time prediction for the editable future time period based at least on the evaluation of the query over the historical data for the one or more past time periods.
2. The method of claim 1 , further comprising at least one of:
weighting at least a portion of the editable future time period and the one or more past time periods; or
weighting at least a portion of the query.
3. The method claim 1 , further comprising:
applying at least one edit to the query, the editable future time period, or the one or more past time periods; and
regenerating the real time prediction of the advertising audience volume based in part or in whole upon the at least one edit to the query, the editable future time period, or the one or more past time periods.
4. The method of claim 1 , wherein generating the real time prediction for the editable future time period further includes:
determining a plurality of weights for the historical advertising audience volumes based at least on recency of data in the historical advertising audience volumes; and
determining the real time prediction for the editable future time period based on applying the plurality of weights to the historical advertising audience volumes.
5. The method of claim 1 , wherein generating the real time prediction for the editable future time period further includes:
determining a plurality of weights for the historical advertising audience volumes based at least on recency of data in the historical advertising audience volumes; and
determining the real time prediction for the editable future time period based on applying the plurality of weights to the historical advertising audience volumes in at least one of a mathematical calculation and a logic determination.
6. The method of claim 1 , further comprising:
applying at least one edit to the query, wherein the at least one edit to the query includes a tuning of at least one category of data included in the query.
7. The method of claim 1 , further comprising:
applying at least one edit to the query, wherein the at least one edit to the query includes a tuning of at least one category of data included in the query, wherein the at least one category is associated with a category type that is at least one of in-market type, demographic type, location type and season type.
8. The method of claim 1 , wherein the real time prediction of the advertising audience volume includes at least a number of persons in the predicted advertising audience volume or a range of the number of persons in the predicted advertising audience volume.
9. The method of claim 1 , wherein changing at least the sampling rate for sampling the historical advertising audience volumes to satisfy the specified level of service and quality of service further comprises determining a subset of the historical data to be sampled based at least on meeting the specified level of service or the quality of service.
10. The method of claim 1 , further comprising determining a confidence metric for the real time prediction.
11. The method of claim 1 , wherein generating the real time prediction for the editable future time period further includes:
determining at least one subset of the historical advertising audience volumes; and
evaluating the query over the determined at least one subset of the historical advertising audience volumes for the one or more past time periods.
12. One or more processor readable non-transitory computer-readable storage media that includes instructions, which when executed by at least one processor, cause the at least one processor to perform a set of acts, the set of acts comprising:
processing a request for a real time prediction that forecasts the advertising audience volume over an editable future time period, the request including at least a query, wherein the request identifies the editable future time period and wherein the request corresponds to a specified level of service and quality of service that determines a number of prediction servers for processing the request;
managing, at a prediction server, a speed of processing the request at least by storing a portion of historical data in memory and another portion of the historical data in one or more persistent storage devices based in part or in whole upon frequencies of accessing the historical data;
evaluating the query on the historical data over the one or more past time periods, by:
identifying historical advertising audience volumes that would have been reached had the query been executed during the one or more past time periods;
retrieving a set of historical data by sampling the historical advertising audience volumes for reducing an amount of the historical data for processing the request, wherein a level of service and a quality of service are balanced against one another by changing at least a sampling rate for sampling the historical advertising audience volumes and a number of employed servers to satisfy the specified level of service and quality of service; and
performing smoothing on the set of historical data retrieved from the historical advertising audience volumes based at least in part upon one or more data sources for the set of historical data and temporal differences in the one or more past time periods; and
generating the real time prediction for the editable future time period based at least on the evaluation of the query over the historical data for the one or more past time periods.
13. The media of claim 12 , the set of acts further comprising enabling at least a portion of the editable future time period and the one or more past time periods to be weighted.
14. The media of claim 12 , the set of acts further comprising enabling at least a portion of the query to be weighted.
15. The media of claim 12 , wherein generating the real time prediction for the editable future time period further includes:
determining a plurality of weights for the historical advertising audience volumes based at least on recency of data in the historical advertising audience volumes; and
determining the real time prediction for the editable future time period based on applying the plurality of weights to the historical advertising audience volumes.
16. The media of claim 12 , wherein generating the real time prediction for the editable future time period further includes:
determining a plurality of weights for the historical advertising audience volumes based at least on recency of data in the historical advertising audience volumes; and
determining the real time prediction for the editable future time period based on applying the plurality of weights to the historical advertising audience volumes in at least one of a mathematical calculation and a logic determination.
17. The media of claim 12 , the set of acts further comprising:
applying at least one edit to the query, wherein the at least one edit to the query includes a tuning of at least one category of data included in the query.
18. The media of claim 12 , the set of acts further comprising:
applying at least one edit to the query, wherein the at least one edit to the query includes a tuning of at least one category of data included in the query, wherein the at least one category is associated with a category type that is at least one of in-market type, demographic type, location type and season type.
19. The media of claim 12 , wherein the real time prediction of the advertising audience volume includes at least a number of persons in the predicted advertising audience volume or a range of the number of persons in the predicted advertising audience volume.
20. The media of claim 12 , wherein changing at least the sampling rate for sampling the historical advertising audience volumes to satisfy the specified level of service and quality of service further comprises determining a subset of the historical data to be sampled based at least on meeting the specified level of service or the quality of service.
21. The media of claim 12 , further comprising determining a confidence metric for the real time prediction.
22. The media of claim 12 , wherein generating the real time prediction for the editable future time period further includes:
determining at least one subset of the historical advertising audience volumes; and
evaluating the query over the determined at least one subset of the historical advertising audience volumes for the one or more past time periods.
23. A system for generating information regarding an advertising audience volume over a network, comprising:
a server device that performs actions, the actions including:
processing a request for a real time prediction that forecasts the advertising audience volume over an editable future time period, the request including at least a query, wherein the request identifies the editable future time period and wherein the request corresponds to a specified level of service and quality of service that determines a number of prediction servers for processing the request;
managing, at a prediction server, a speed of processing the request at least by storing a portion of historical data in memory and another portion of the historical data in one or more persistent storage devices based in part or in whole upon frequencies of accessing the historical data;
evaluating the query on the historical data over the one or more past time periods, by:
identifying historical advertising audience volumes that would have been reached had the query been executed during the one or more past time periods;
retrieving a set of historical data by sampling the historical advertising audience volumes for reducing an amount of the historical data for processing the request, wherein a level of service and a quality of service are balanced against one another by changing at least a sampling rate for sampling the historical advertising audience volumes and a number of employed servers to satisfy the specified level of service and quality of service; and
performing smoothing on the set of historical data retrieved from the historical advertising audience volumes based at least in part upon one or more data sources for the set of historical data and temporal differences in the one or more past time periods; and
generating the real time prediction for the editable future time period based at least on the evaluation of the query over the historical data for the one or more past time periods.
24. The system of claim 23 , the server device further performing actions including:
enabling at least a portion of the editable future time period and the one or more past time periods to be weighted.
25. The system of claim 23 , the server device further performing actions including:
enabling at least a portion of the query to be weighted.
26. The system of claim 23 , wherein generating the real time prediction for the editable future time period further includes:
determining a plurality of weights for the historical advertising audience volumes based at least on recency of data in the historical advertising audience volumes; and
determining the real time prediction for the editable future time period based on applying the plurality of weights to the historical advertising audience volumes.
27. The system of claim 23 , wherein generating the real time prediction for the editable future time period further includes:
determining a plurality of weights for the historical advertising audience volumes based at least on recency of data in the historical advertising audience volumes; and
determining the real time prediction for the editable future time period based on applying the plurality of weights to the historical advertising audience volumes in at least one of a mathematical calculation and a logic determination.
28. The system of claim 23 , the server device further performing actions including:
applying at least one edit to the query, wherein the at least one edit to the query includes a tuning of at least one category of data included in the query.
29. The system of claim 23 , the server device further performing actions including:
applying at least one edit to the query, wherein the at least one edit to the query includes a tuning of at least one category of data included in the query, wherein the at least one category is associated with a category type that is at least one of in-market type, demographic type, location type and season type.
30. The system of claim 23 , wherein the real time prediction of the advertising audience volume includes at least one of a number of persons in the predicted advertising audience volume or a range of the number of persons in the predicted advertising audience volume.
31. The system of claim 23 , wherein changing at least the sampling rate for sampling the historical advertising audience volumes to satisfy the specified level of service and quality of service further comprises determining a subset of the historical data to be sampled based at least on meeting the specified level of service or the quality of service.
32. The system of claim 23 , further comprising determining a confidence metric for the real time prediction.
33. The system of claim 23 , wherein generating the real time prediction for the editable future time period further includes:
determining at least one subset of the historical advertising audience volumes; and
evaluating the query over the determined at least one subset of the historical advertising audience volumes for the one or more past time periods.
34. An apparatus for generating information regarding an advertising audience volume over a network, comprising:
a memory device for storing data and instructions; and
processor device that is configured to execute instructions stored in the memory device, wherein executions of the instructions by the processor device cause the processor device to:
process a request for a real time prediction that forecasts the advertising audience volume over an editable future time period, the request including at least a query, wherein the request identifies the editable future time period and wherein the request corresponds to a specified level of service and quality of service that determines a number of prediction servers for processing the request;
manage, at a prediction server, a speed of processing the request at least by storing a portion of historical data in memory and another portion of the historical data in one or more persistent storage devices based in part or in whole upon frequencies of accessing the historical data;
evaluate the query on the historical data over the one or more past time periods, by:
identifying historical advertising audience volumes that would have been reached had the query been executed during the one or more past time periods;
retrieving a set of historical data by sampling the historical advertising audience volumes for reducing an amount of the historical data for processing the request, wherein a level of service and a quality of service are balanced against one another by changing at least a sampling rate for sampling the historical advertising audience volumes and a number of employed servers to satisfy the specified level of service and quality of service; and
perform smoothing on the set of historical data retrieved from the historical advertising audience volumes based at least in part upon one or more data sources for the set of historical data and temporal differences in the one or more past time periods; and
generate the real time prediction for the editable future time period based at least on the evaluation of the query over the historical data for the one or more past time periods.
35. The apparatus of claim 34 , wherein executions of the instructions by the processor device further cause the processor device to:
enable at least a portion of the editable future time period and the one or more past time periods to be weighted.
36. The apparatus of claim 34 , wherein executions of the instructions by the processor device further cause the processor device to:
enable at least a portion of the query to be weighted.
37. The apparatus of claim 34 , wherein generating the real time prediction for the editable future time period further includes:
determining a plurality of weights for the historical advertising audience volumes based at least on recency of data in the historical advertising audience volumes; and
determining the real time prediction for the editable future time period based on applying the plurality of weights to the historical advertising audience volumes.
38. The apparatus of claim 34 , wherein generating the real time prediction for the editable future time period further includes:
determining a plurality of weights for the historical advertising audience volumes based at least on recency of data in the historical advertising audience volumes; and
determining the real time prediction for the editable future time period based on applying the plurality of weights to the historical advertising audience volumes in at least one of a mathematical calculation and a logic determination.
39. The apparatus of claim 34 , wherein executions of the instructions by the processor device further cause the processor device to:
apply at least one edit to the query, wherein the at least one edit to the query includes a tuning of at least one category of data included in the query.
40. The apparatus of claim 34 , wherein executions of the instructions by the processor device further cause the processor device to:
apply at least one edit to the query, wherein the at least one edit to the query includes a tuning of at least one category of data included in the query, wherein the at least one category is associated with a category type that is at least one of in-market type, demographic type, location type and season type.
41. The apparatus of claim 34 , wherein the real time prediction of the advertising audience volume includes at least one of a number of persons in the predicted advertising audience volume or a range of the number of persons in the predicted advertising audience volume.
42. The apparatus of claim 34 , wherein changing at least the sampling rate for sampling the historical advertising audience volumes to satisfy the specified level of service and quality of service further comprises determining a subset of the historical data to be sampled based at least on meeting the specified level of service or the quality of service.
43. The apparatus of claim 34 , wherein executions of the instructions by the processor device further cause the processor device to:
determine a confidence metric for the real time prediction.
44. The apparatus of claim 34 , wherein generating the real time prediction for the editable future time period further includes:
determining at least one subset of the historical advertising audience volumes; and
evaluating the query over the determined at least one subset of the historical advertising audience volumes for the one or more past time periods.