Systems and methods for adaptively identifying and mitigating statistical outliers in aggregated data

• US 10,417,201 B2
• Filed: 08/13/2015
• Issued: 09/17/2019
• Est. Priority Date: 08/13/2014
• Status: Active Grant

• Current Assignee: FINCA International, Inc.
• Original Assignee: FINCA International, Inc.
• Inventors: Graham, Scott, Tevosyan, Anahit
• Primary Examiner(s): Nguyen, Cam Linh T
• Application Number:
  US14/826,019
• Publication Number:
  US 20160048552A1
• Time in Patent Office: 1,496 Days
• Field of Search: 707690
• CPC Class Codes: G06F 16/215 Improving data quality; Data cleansing, e.g. de-duplication, removing invalid entries or correcting typographical errors
• CPC Class Codes: G06Q 30/02 Marketing; Price estimation or determination; Fundraising

Abstract

The disclosed embodiments include computerized methods and systems that facilitate automated detection and precision correction of aggregated data collected by multiple, geographically dispersed mobile communications devices. In one embodiment, an apparatus detect a data outlier within portions of the aggregated data having numerical and/or categorical values. The apparatus may transmit information identifying the data outliner and a portion of the aggregated data that includes the data outlier to an additional communications device, which may present the aggregated data portion to a user in a manner that visually distinguishes the data outlined from other elements of aggregated data. In response to a request from the additional communications device, the apparatus may modify portions of the aggregated data in an effort to mitigate the data outlier.

Claims

1. An apparatus, comprising:

a storage device; and

at least one processor coupled to the storage device, the storage device storing software instructions that are executable by the at least one processor, and the at least one processor being operative with the software instructions and configured to;

obtain aggregated data collected by a plurality of first communications devices associated with corresponding surveyors, the aggregated data comprising aggregated survey response data collected by the first communications devices;

detect at least one data outlier within the aggregated survey response data by applying a numerical technique comprising at least one of a linear regression analysis, conditional or robust regressions, an interquartile outlier check, or a standard deviation analysis, the data outlier corresponding to at least one element of the aggregated survey response data;

determine a magnitude by which the at least one data outlier exceeds a validation limit based on results from the numerical technique;

transmit information identifying the data outlier and at least a portion of the aggregated data that includes the data outlier to a second communications device associated with a survey manager, the information instructing the second communications device to;

generate a first graphical user interface displaying the portion of the aggregated survey response data that includes the data outlier, the data outlier being displayed visually distinguishable within the presented aggregated survey response data portion and being associated with a hyperlink, and

generate a second graphical user interface when the survey manager selects the hyperlink, the second graphical user interface comprising an editable box and displaying the magnitude by which the data outlier exceeds a validation limit, the second graphical user interface being different from the first graphical user interface;

in response to the transmitted information, receive a request to modify the aggregated data from the second communications device, the request comprising a value entered in the editable box;

perform operations that modify at least a portion of the aggregated data in accordance with the received request; and

generate metadata associated with the aggregated data, the metadata comprising;

effected modifications to the portion of the aggregated data; and

sources of the effected modifications.

2. The apparatus of claim 1, wherein the aggregated data comprises a first data portion and a second data portion, the first data portion including numerical values, and the second data portion including categorical data, and the categorical data including at least one non-numerical value.

3. The apparatus of claim 2, wherein the at least one processor is further configured to:

apply the numerical technique to a subset of the numerical data, the numerical data subset being characterized by a data parameter, the numerical technique being based on a corresponding data type of the numerical data; and

identify the data outlier based on an outcome of the numerical technique.

4. The apparatus of claim 3, wherein the at least one processor is further configured to:

obtain information specifying a validation rule, the validation rule defining a valid relationship between elements of the aggregated data and the outcome of the numerical technique;

determine that at least one element of the aggregated data fails to exhibit the valid relationship with the outcome of the numerical technique; and

establish the at least one aggregated survey response data element as the data outlier.

5. The apparatus of claim 2, wherein the at least one processor is further configured to:

determine a probability of an occurrence of a first element of the categorical data in conjunction with at least one second element of the categorical data;

determine that the probability fails to exceed at threshold value; and

establish the first element of categorical data as the data outlier in response to the determination.

6. The apparatus of claim 2, wherein the at least one processor is configured to perform operations that normalize at least a portion of the numerical data in accordance with one or more normalization criteria, the normalization criteria being based on a number of household occupants.

7. The apparatus of claim 1, wherein the at least one processor is further configured to:

identify a first portion of the aggregated survey response data collected by a corresponding one of the first communications devices;

determine a metric indicative of a degree of separation between the first portion of the aggregate data and other portions of aggregated survey response data; and

based on the determined metric, establish that at least one element of the first aggregated survey response data portion represents potential surveyor falsified data.

8. The apparatus of claim 1, wherein the at least one processor is further configured to:

detect a plurality of data outliers within the aggregated survey response data; and

transmit information identifying the plurality of data outliers and at least a portion of the aggregated survey response data that includes the data outliers to the second communications device, the information identifying the plurality of data outliers comprising instructions to instruct the second communications device to display a list of identifiers of the plurality of first communications devices that are associated with a list of query identifiers.

9. The apparatus of claim 1, wherein the at least one processor is configured to re-format at least a portion of the aggregated survey response data in accordance with a predetermined data format.

10. The apparatus of claim 1, wherein

the metadata further comprises one or more of:

information of surveyors associated with the portion of the aggregated data;

locations associated with the portion of the aggregated data;

survey questions associated with the portion of the aggregated data; and

survey client identifiers associated with the portion of the aggregated data;

the second graphical user interface further comprises;

a normalized value displayed between the editable box and the magnitude by which the data outlier exceeds the validation limit; and

the first graphical interface comprises;

a plurality of query identifiers for corresponding survey indices, each query identifier comprising raw response values and normalized response values.

11. The apparatus of claim 1, wherein

the surveyors are users of the plurality of first communication devices;

the survey manager is a user of the second communication device; and

the survey manager assigns surveyors one or more surveys to collect and process survey data.

12. A computer-implemented method, comprising:

obtaining, by at least one processor, aggregated data collected by a plurality of first communications devices associated with corresponding surveyors, the aggregated data comprising aggregated survey response data collected by the first communications devices;

detecting, by the at least one processor, at least one data outlier within the aggregated survey response data by applying a numerical technique comprising at least one of a linear regression analysis, conditional or robust regressions, an interquartile outlier check, or a standard deviation analysis, the data outlier corresponding to at least one element of the aggregated survey response data;

determining a magnitude by which the at least one data outlier exceeds a validation limit based on results from the numerical technique;

generating, by the at least one processor, an electronic command to transmit information identifying the data outlier and at least a portion of the aggregated data that includes the data outlier to a second communications device associated with a survey manager, the information instructing the second communications device to;

generate a first graphical user interface displaying the portion of the aggregated survey response data that includes the data outlier, the data outlier being displayed visually distinguishable within the presented aggregated survey response data portion, and being associated with a hyperlink, and

generate a second graphical user interface when the survey manager selects the hyperlink, the second graphical user interface comprising an editable box and displaying the magnitude by which the data outlier exceeds a validation limit, the second graphical user interface being different from the first graphical user interface;

in response to the transmitted information, receiving, by at least one processor, a request to modify the aggregated data from the second communications device, the request comprising a value entered in the editable box;

performing, by at least one processor, operations that modify at least a portion of the aggregated data in accordance with the received request; and

generating metadata associated with the aggregated data, the metadata comprising;

effected modifications to the portion of the aggregated data; and

sources of the effected modifications.

13. The method of claim 12, wherein the aggregated data comprises a first data portion and a second data portion, the first data portion including numerical values, and the second data portion including categorical data, and the categorical data including at least one non-numerical value.

14. The method of claim 13, further comprising:

applying numerical technique to a subset of the numerical data, the numerical data subset being characterized by a data parameter, the numerical technique being based on a corresponding data type of the numerical data; and

identifying the data outlier based on an outcome of the numerical technique.

15. The method of claim 14, further comprising:

obtaining information specifying a validation rule, the validation rule defining a valid relationship between elements of the aggregated data and the outcome of the numerical technique;

determining that at least one element of the aggregated data fails to exhibit the valid relationship with the outcome of the numerical technique; and

establishing the at least one aggregated data element as the data outlier.

16. The method of claim 13, further comprising:

determining a probability of an occurrence of a first element of the categorical data in conjunction with at least one second element of the categorical data;

determining that the probability fails to exceed at threshold value; and

establishing the first element of categorical data as the data outlier in response to the determination.

17. The method of claim 13, further comprising

performing operations that normalize at least a portion of the numerical data in accordance with one or more normalization criteria, the normalization criteria being based on a number of household occupants.

18. The method of claim 12, further comprising:

identifying a first portion of the aggregated survey response data collected by a corresponding one of the first communications devices;

determining a metric indicative of a degree of separation between the first portion of the aggregate data and other portions of aggregated survey response data; and

based on the determined metric, establishing that at least one element of the first aggregated survey response data portion represent potential surveyor falsified data.

19. The method of claim 12, further comprising:

detect a plurality of data outliers within the aggregated survey response data; and

transmit information identifying the plurality of data outliers and at least a portion of the aggregated survey response data that includes the data outliers to the second communications device, the information identifying the plurality of data outliers comprising instructions to instruct the second communications device to display a list of identifiers of the plurality of first communications devices that are associated with a list of query identifiers.

20. A tangible, non-transitory computer-readable medium storing instructions that, when executed by at least one processor, perform a method comprising:

obtaining aggregated data collected by a plurality of first communications devices associated with corresponding surveyors, the aggregated data comprising aggregated survey response data collected by the first communications devices;

detecting at least one data outlier within the aggregated survey response data by applying a numerical technique comprising at least one of a linear regression analysis, conditional or robust regression, and an interquartile check, or a standard deviation analysis, the data outlier corresponding to at least one element of the aggregated survey response data;

determining a magnitude by which the at least one data outlier exceeds a validation limit based on results from the numerical technique;

generating an electronic command to transmit information identifying the data outlier and at least a portion of the aggregated data that includes the data outlier to a second communications device associated with a survey manager, the information instructing the second communications device to;

generate a first graphical user interface displaying the portion of the aggregated survey response data that includes the data outlier, the data outlier being displayed visually distinguishable within the presented aggregated survey response data portion, and being associated with a hyperlink, and

generate a second graphical user interface when the survey manger selects the hyperlink, the second graphical user interface comprising an editable box and displaying the magnitude by which the data outlier exceeds a validation limit, the second graphical user interface being different from the first graphical user interface;

in response to the transmitted information, receiving a request to modify the aggregated data from the second communications device, the request comprising a value entered in the editable box;

performing operations that modify at least a portion of the aggregated data in accordance with the received request; and

generating metadata associated with the aggregated data, the metadata comprising;

effected modifications to the portion of the aggregated data; and

sources of the effected modifications.