System and method for cleaning television viewing measurement data

A computer-implemented method for cleaning television viewing data by detecting and compensating periodic non-human generated tuning events, the method comprising:

• accessing set top box viewing data, the television viewing data comprised of a plurality of tuning events;
• identifying tunings events in the accessed television viewing data that exhibit a periodicity to a channel over a time period in excess of a threshold;
• categorizing tuning events exhibiting a periodicity to a channel over a time period in excess of a threshold as non-human generated tuning events;
• and
• compensating for the non-human generated tuning events in the television viewing data.

A computer-implemented method for cleaning television viewing data by detecting and compensating for non-human generated tuning events in groups of set top boxes, the method comprising:

• selecting television viewing data associated with a plurality of set top boxes, the television viewing data comprised of a plurality of tuning events;
• analyzing the television data to identify a tuning event occurring at a same time for each set top box within the plurality of set top boxes;
• characterizing the identified tuning event as non-human generated;
• and
• compensating for the non-human generated tuning event in the television viewing data.

A computer-implemented method for cleaning television viewing data by detecting and compensating for clock offsets, the method comprising:

• accessing television viewing data associated with a first plurality of set top boxes and television viewing data associated with a second plurality of set top boxes;
• determining a first viewing profile corresponding to the viewing data associated with the first plurality of set top boxes;
• determining a second viewing profile corresponding to the viewing data associated with the second plurality of set top boxes;
• determining a time offset that when added to the first viewing profile improves correlation with the second viewing profile;
• and
• compensating for a clock offset between the first and second pluralities of set top boxes based on a relation between the determined time offset and a threshold.

A system for cleaning television viewing data so it can be analyzed with improved accuracy, comprising:

• at least one processor;
• at least one memory storing computer-executable instructions configured to cause the at least one processor to:
• access television viewing data that is associated with a plurality of set top boxes,
• clean the accessed television viewing data, wherein cleaning the accessed television viewing data comprises:
• compensating for buffer overflow within a set top box;
• compensating for non-human generated tunes captured by a set top box;
• compensating for clock offsets between two pluralities of set top boxes, and
• analyzing at least a portion of the television viewing data after it is cleaned.