JP6201077B1 - Investigation data processing apparatus and investigation data processing method - Google Patents

Abstract

Even if one or both of survey data indicating the results of surveys of a first survey and a second survey conducted on different monitors are dynamically changed, the response contents can be accurately and accurately determined. Provide data processing services for efficient fusion and effective use as fused data. New first survey data is identified among first survey data (S201). A distance calculation score group is calculated by applying a function to the common items of both the first survey data and the second survey data (S202). The distance calculation score groups are compared, and distance calculation is executed for the distance indicating the degree of similarity between each monitor A 'of the new first survey data and each monitor B of the second survey data (S203). For each new monitor A ′ of the first survey data, a monitor having a short total distance is identified from the monitors B of the second survey data, and they are merged by regarding them as the same monitor (S204). The fusion data is saved (S205). [Selection] Figure 12

Description

  The present invention relates to a survey data processing apparatus and a survey data processing method, and more particularly to an apparatus and method for processing survey data indicating the results of surveys of a first survey and a second survey performed on different monitors.

  Various surveys such as questionnaire surveys are conducted for the purpose of media planning, etc., and survey data indicating the survey results are often processed in each industry. For example, each company has its own products and services. Conduct a survey and extract useful information from the survey results. Furthermore, some surveys are conducted for the purpose of collecting multifaceted information such as purchase, advertisement contact, lifestyle, etc. for the same monitor. This is referred to as “single source data” and is a useful information source when analyzing the correlation between advertising media and whether or not a product is purchased on an individual basis.

  On the other hand, a technique for fusing a plurality of databases is conventionally known. Such a technique is called data fusion, and is used as a means for solving the disadvantage that it is very expensive to acquire single source data.

  Referring to the conceptual diagram of data fusion shown in FIG. 15, two survey data A and B acquired independently from each other are fused based on the similarity of the monitors that have answered each survey. In other words, the monitors having a high degree of similarity with respect to the answer contents of the common items 1 to n are regarded as the same monitor, and thereby the record indicating the answer contents of a certain monitor in the survey data A (specifically, the common items 1 to n and It is possible to give a record (specifically, the answer of the original items B1 to n) indicating the response contents of the monitor that is regarded as the certain monitor in the survey data B to the answer of the original items A1 to n). It becomes possible. As a result, fusion data is generated by adding the response contents of the original items B1 to n not recorded in the survey data A to the contents of the survey data A.

  As described above, even the survey results set to a relatively small number of questions by data fusion can be converted to single source data, and the pseudo single source data obtained in this way can be used. Thus, information useful in media planning and the like can be obtained at a lower cost.

  Further, as data fusion methods, “unconstrained statistical matching” and “constrained statistical matching” can be cited. “Unconstrained statistical matching” simply fuses data based on the similarity of monitors. This “unconstrained statistical matching” has an advantage that a large amount of data can be processed efficiently because the processing is simplified. On the other hand, “Statistical matching with constraints” fuses data under the constraint of maintaining the average and variance of the original data, so the information that the original data has can be transferred to the fusion data more accurately. is there.

  In addition, with “constrained statistical matching”, the time required for calculation for data fusion usually takes a long time. In recent years, the calculation time has been shortened and accuracy that is almost the same as “constrained” has been verified. A technique called “statistical matching with distance priority constraint” has been developed (see Patent Document 1).

Japanese Patent No. 4338486

As described above, data fusion is effective as a method for obtaining single source data from the survey results set in a relatively small number of questions.
In other words, by effectively utilizing the fusion data obtained by fusing the survey results of a plurality of surveys, it is possible to perform aggregation that is impossible before the fusion. Further, if execution of a series of processes related to this is provided as a service by an ASP (Application Service Provider) or the like, it will be beneficial for each company that wants to use the survey results.

However, in the case of survey data in which one or both of the survey data changes dynamically, records that could be merged at the time of data fusion no longer exist in the survey data, or new records that did not exist at the time of data fusion. Records may appear later.
For example, as an example of the survey data that dynamically changes in this way, panel survey data for rotation is given. Specifically, long-term investigations of the same subject in a panel survey may have an impact due to survey familiarity and learning effects, and the sample may not respond to changes in the population. In order to avoid such a situation, in the panel survey, a part or all of the monitor may be sequentially rearranged (rotation such as total replacement or partial replacement) after a certain period of time. In addition to the periodic rotation as described above, the panel survey data may be replaced irregularly by adding a new monitor to supplement the dropped monitor. Furthermore, not only survey data such as questionnaire surveys as described above, but dynamically changing survey data includes operation log data that can be automatically acquired from various devices, web logs, purchase history, application usage history, etc. Alternatively, a large amount of customer data that changes constantly (so-called big data) can also be considered as dynamically changing survey data.

  As described above, there are various methods for data fusion, such as “unconstrained statistical matching”, “constrained statistical matching”, or “statistic matching with distance priority constraint”. In the survey data in which one or both of the survey data changes dynamically, it has not been possible to accurately and efficiently fuse new records that did not exist at the time of data fusion.

  Therefore, the present invention has been made in view of the above-mentioned problems, and the object thereof is one or both of the survey data indicating the survey results of the first survey and the second survey conducted on different monitors. Survey data processing apparatus capable of providing a data processing service for accurately and efficiently merging the contents of responses even if the survey data is dynamically changing It is to realize the survey data processing method.

  According to the survey data processing device of the present invention, the problem is a survey data processing device that processes survey data indicating the results of each of the first survey and the second survey performed on different monitors, A first survey data acquisition unit that acquires first survey data collected for the number of persons of the first survey monitor who answered the first survey for the contents of responses to common items included in both the first survey and the second survey And second survey data obtained by collecting the number of responses of the second survey monitor who answered the second survey, and the contents of the responses to the common survey items and the second survey unique items included only in the second survey. The second survey data acquisition unit, the first survey data acquired by the first survey data acquisition unit, and the second survey data acquired by the second survey data acquisition unit, 1 Survey data And a process execution unit for executing a data fusion process for generating fused data indicating the response contents of each of the second survey data, and the data fusion process is performed on the common item using a predetermined arithmetic expression. A calculation process for calculating a value related to the response content, and using the result of comparing the values between the first survey monitor and the second survey monitor to calculate the similarity of the response content to the common item; An assignment process for assigning, to the first survey monitor, the same response content as the response content to the second survey original item of the second survey monitor in the allocation pattern set based on the calculated similarity degree; And when both or one of the first survey monitor and the second survey monitor has changed, the allocation process has already been performed. For the first survey monitor and the second survey monitor that still exist at that time, the allocation partner is taken over as it is, and the calculation process and the allocation are performed only for the new first survey monitor or the new second survey monitor. It is solved by executing the process.

  Further, according to the survey data processing method of the present invention, a survey data processing method for processing survey data indicating the survey results of the first survey and the second survey performed on different monitors by a computer, The computer acquiring first survey data in which the contents of responses to common items included in both the first survey and the second survey are collected for the number of first survey monitors who answered the first survey; The second survey data obtained by collecting the contents of the responses to the second survey unique items included in the common survey and the second survey only for the number of the second survey monitors who answered the second survey is acquired. The process and the acquired first survey data and the second survey data are merged to generate fused data indicating the contents of the responses of the first survey data and the second survey data. Performing a data fusion process, wherein the data fusion process calculates a value related to a response content for the common item using a predetermined arithmetic expression, and the first survey monitor and the second survey monitor In the calculation process for calculating the degree of similarity of the answer contents for the common item using the result of comparing the values with each other, and the assignment pattern set based on the calculated degree of similarity, the first An assignment process for assigning to the survey monitor the same response content as the response content to the second survey original item of the second survey monitor, and either or both of the first survey monitor and the second survey monitor Change, the first investigation monitor and the second monitor that have already been allocated by the allocation process and the allocation partner still exists at that time. Take over as assignment partner for 査 monitor, performing the calculation process and the assignment process only the new first study monitor or the new second study monitor is solved by.

In the survey data processing apparatus or the survey data processing method of the present invention configured as described above, when both or one of the first survey monitor and the second survey monitor changes, the allocation has already been performed by the allocation process. For the first survey monitor and the second survey monitor that still exist at that time, the allocation partner is taken over as it is, and the calculation process and the allocation process are executed again only for the new first survey monitor or the new second survey monitor.
As a result, even if one or both of the survey data indicating the survey results of the first survey and the second survey performed on different monitors are dynamically changed, the fused data that has already been fused. Can be utilized as it is, and the response contents can be continuously and accurately fused. Then, by analyzing the information indicated by the fusion data, it is possible to obtain information that was not obtained from the information before the fusion, and to plan effective media planning based on the information.

  Further, when a preferred configuration of the survey data processing apparatus is described, the predetermined arithmetic expression used in the calculation process is changed before and after either or both of the first survey monitor and the second survey monitor are changed. Do not do it.

  As in the above configuration, when merging new survey data, a predetermined arithmetic expression (function such as a fusion parameter) used to calculate a value related to the response content (distance calculation score) for the common item is re-executed. By reusing a predetermined arithmetic expression that has already been used without calculation, the processing is simplified, and even a large amount of new survey data can be processed quickly.

  A preferred configuration of the survey data processing apparatus described above is as follows. When the survey time for the common item is different before and after both or one of the first survey monitor and the second survey monitor is changed, the calculation process is performed. In the assignment process, the process execution unit may substitute the response content for the item or time approximate to the common item or the statistical value of the response content as the response content for the common item.

  When the new survey data does not hold the same common item as the fusion data due to the difference in the survey period for the common item, for example, because the common item depends on the time point, as in the above configuration Even so, it is possible to maintain the same common items by substituting the answer contents for the items or time approximate to the common items or the statistical values such as the average value or representative value of the answer contents as the answer contents for the common items. New survey data can be merged without any difference.

In addition, when a preferred configuration is described for the survey data processing apparatus, the process execution unit performs the first process in the allocation process before either or one of the first survey monitor and the second survey monitor changes. 1 survey totalization result is aligned between the fusion data and the first survey data, and the like aligned between the second regulating counting result about the査 said second survey data and the fusion data The allocation pattern should be set according to a statistical solution.

  As described above, before both or one of the first survey monitor and the second survey monitor change, so-called “constrained statistical matching” is adopted, and the aggregated results in the survey data before merging are the surveys after merging. By performing the fusion process so that it is aligned with the aggregated results in the data (fusion data), the average and variance of the first survey data and the second survey data can be maintained, and the information held by each survey data is more accurately fused data. Can take over.

  Moreover, when a suitable structure is described about said investigation data processing apparatus, it is good for the said 1st investigation data to contain the reply content with respect to the 1st investigation original item contained only in the said 1st investigation.

  As in the above configuration, the first survey data includes the answer contents for the first survey unique items included only in the first survey in addition to the common items, so that it is possible to obtain fusion data with more complete content.

  The preferred configuration of the survey data processing apparatus is as follows. The first survey is a survey on media contact, and the second survey is a questionnaire that captures various aspects of consumer attributes, product involvement, and media contact. It is good to be a survey.

  As in the above configuration, the survey data on one side of the TV program, radio program, etc. (including CM), information that can be viewed and viewed on the Internet by mobile phone, smartphone, tablet, PC, etc. (website, etc.) Surveys related to media contact such as browsing status, viewing status of videos and music, subscription status of newspapers and magazines, etc., and other survey data to be a questionnaire survey that captures multifaceted aspects of consumer attributes, product involvement, media contact, etc. Therefore, it is possible to grasp these correlations, and it is possible to select an advertising space (broadcast time, TV program, etc.) that can obtain the maximum advertising effect based on this correlation. Can be used more effectively.

According to the present invention, even if one or both of the survey data indicating the survey results of the first survey and the second survey conducted on different monitors are dynamically changed, the response contents It is possible to realize a survey data processing apparatus and a survey data processing method capable of providing a data processing service for accurately and efficiently merging the data and utilizing them effectively as fused data.
Thus, by analyzing the information indicated by the fusion data, it is possible to obtain information that could not be obtained from the information before the fusion, and to plan effective media planning based on the information. More specifically, for example, grasping the correlation between items surveyed in one survey (for example, whether or not watching TV) and items surveyed in other surveys (for example, living conditions and consumer awareness) It is possible to select an advertisement frame (broadcast time, television program, etc.) that provides the maximum advertising effect based on this correlation.

It is explanatory drawing regarding the data processing service which concerns on this invention. It is the figure which showed the communication system containing the investigation data processing apparatus of this invention. (A) in FIG. 3 shows the answer results for the first survey, and (B) shows the answer results for the second survey. It is the figure which showed the hardware constitutions of the investigation data processing apparatus. It is the figure which showed the structure of the investigation data processing apparatus from the functional surface. It is the figure which showed the content of the data fusion process notionally. It is the figure which showed the rough flow about the series of processes which concern on a data processing service. It is a figure which shows the flow of the data fusion process in the time t. It is explanatory drawing at the time of employ | adopting "unconstrained statistical matching" and implementing a data fusion process. It is explanatory drawing at the time of employ | adopting "restricted statistical matching" and performing a data fusion process. It is a figure which shows notionally the state to which the 1st investigation data changed. It is a figure which shows the flow of the data fusion process in the time t + 1 after fusion implementation. It is a figure which shows notionally the state to which the 1st investigation data changed. It is a figure which shows notionally the state to which the 1st survey data changed after the 2nd survey data changed. It is the figure which showed the concept of data fusion.

Hereinafter, an embodiment of the present invention (hereinafter, this embodiment) will be described with reference to the drawings.
In this embodiment, an audience rating survey, which is an example of a media contact survey, is a first survey, and a questionnaire survey that captures multifaceted aspects of consumer attributes, product involvement, media contacts, and the like is a second survey. I will give you a description. Here, in the relationship between the two surveys, the audience rating survey is the main and the questionnaire survey is the subordinate.
In the following description, “single source data” is data indicating the contents of answers to a relatively large number of questions (for example, several hundred questions) collected from the same monitor (respondent). This is data that collects multifaceted information such as lifestyle. Specifically, the information indicated by the data includes the contents of answers to questions regarding monitor attributes, specifically information related to demographics (demographic attributes) and psychographics (psychological attributes). Information etc. are included.
Note that the survey on media contact is not limited to the survey on the viewing status of television programs (including CMs) as in this embodiment, but the viewing status of radio programs (including CMs), mobile phones, smartphones, tablets, PCs, etc. The browsing status of information (website etc.) that can be browsed / viewed on the Internet, viewing status of videos / music, subscription status of newspapers / magazines, etc. may be used.

<< About Survey Data Processing Service >>
First, a survey data processing service realized by the present invention will be described with reference to FIG. FIG. 1 is an explanatory diagram relating to a survey data processing service.

  The survey data processing service is provided by a survey company that manages the survey data processing apparatus of the present invention, and a plurality of surveys performed on different monitors (in this embodiment, the first survey and the second survey). It is a service that executes a series of data processing for fusing survey data indicating the survey results of the two types of surveys.

More specifically, the research company conducts an audience rating survey as the first survey. This audience rating survey is a survey on the viewing status of TV programs, etc., on randomly selected monitors. When the survey company collects information on the viewing status from each monitor, the information is collected. Are collected and stored as the first survey data.
Note that there are two types of audience ratings: household audience ratings and individual audience ratings. In this embodiment, the audience ratings are used. However, the present invention is not limited to this, and it can be applied even when the audience rating is the household to be surveyed.
The individual audience rating is the number of members who belong to a household, and indicates who has watched what kind of TV program and how many of them are in a family of a certain age or older (for example, 4 years or older) in the household. It is a ratio, and is used when it is desired to know how much an individual with what attribute has watched by dividing the viewer into gender as in this embodiment. On the other hand, the household audience rating is a ratio indicating how many of the TV-owned households are watching TV programs, with the survey target as household units.

In addition, the survey company conducts a questionnaire survey as the second survey. In this questionnaire survey, a plurality of questions are asked for a plurality of monitors. When the survey company collects answers to all the questions from each monitor, the survey company collects the answers corresponding to the number of monitors and sets the second answer. Store as survey data.
In the questionnaire survey, three viewpoints of consumer attributes, product involvement, and media contact are surveyed on the same monitor, and a relatively large number of questions are given to each monitor for the purpose of grasping consumers from various perspectives. Specifically, there are about several hundred questions, and the contents include demographic attributes and psychographic attributes of the monitor. That is, the survey data indicating the result of the questionnaire survey conducted by the survey company as the second survey corresponds to the single source data.
The questionnaire survey monitor is a different monitor from the audience rating survey monitor. In addition, the number of monitors for which the survey company requests a questionnaire survey can be set freely, but it is desirable to secure monitors on a scale that can collect sufficient information as single source data.

The research company merges the first survey data acquired in the audience rating survey as the first survey with the second survey data as single source data acquired in the questionnaire survey as the second survey. As a result, fusion data obtained by fusing the two survey data is generated.
As described above, the research company can execute the data fusion process, aggregate the response results of the first survey and the second survey, and obtain the result in the form of fusion data, which is impossible with the data before fusion. It is also possible to perform total tabulation. In addition, this fusion data can be provided to client companies such as manufacturers, and the customer company receiving the fusion data analyzes the information indicated by the fusion data, and performs effective media planning based on the information. Can be planned.

<< About survey data processing service provision system >>
Next, a system configuration for providing the above-described data processing service will be described with reference to FIG. FIG. 2 is a diagram showing a communication system including the survey data processing apparatus of the present invention.
In order to provide the above-described data processing service, the research company has a computer, more precisely, a server computer (hereinafter referred to as server 1). The server 1 corresponds to the survey data processing apparatus of the present invention, and provides the above-described data processing service as an ASP service.

  When the function of the server 1 is outlined, as shown in FIG. 2, the server 1 is communicably connected to each household subject to audience rating survey via an information communication network N1 such as the Internet.

Specifically, the survey company distributes the measuring device 21 to the extracted household. This measuring device 21 is placed in the home of the extracted household (strictly, around the location where the TV is installed) and measures the viewing status of the TV program with each member in the entire household as one survey subject. Used to do. Data indicating the viewing status of the television program generated by the measuring device 21 is transmitted to an external device through the information communication network N1. The data (hereinafter referred to as audience rating data) transmitted by the measuring device 21 is data indicating which television program is viewed by which survey target and how much.
Incidentally, in this embodiment, the members of the household (that is, individuals) are the survey subjects, but the household itself may be handled as the survey subjects.

  As shown in FIG. 2, the server 1 is communicably connected to an answer terminal 31 owned by a monitor via an information communication network N2 such as the Internet. Then, the server 1 communicates with each answering terminal 31 to receive answer data (hereinafter, individual answer data) of each monitor for each question of the questionnaire survey.

Here, the answering terminal 31 is a tablet-type communication terminal distributed by a research company to each monitor that has requested a questionnaire survey response. In other words, the monitor looks at the touch panel mounted on the answering terminal 31 to confirm each question of the questionnaire survey, and answers through a touch operation on the touch panel. When the answer terminal 31 accepts an answer operation by the monitor, individual answer data indicating the answer content is generated, and the generated data is transmitted to the server 1.
The answering terminal 31 is not limited to a tablet-type terminal, and may be another communication terminal such as a smartphone or a notebook PC. Further, instead of the answering terminal 31, a method may be used in which the survey company distributes questionnaire sheets to each monitor, and the survey company side converts the questionnaire data into data based on the contents of the collected questionnaire sheets that have been answered.

  In addition, when the server 1 receives the number of individual response data corresponding to the specified number of questions from each response terminal 31, the information indicated by each data is collected and aggregated as response results for all monitors, and the aggregated data That is, the second survey data is stored in the server 1.

In addition, the server 1 executes a process of generating the fusion data by fusing the first investigation data and the second investigation data stored in the server 1, that is, a data fusion process. Here, the generation of the fusion data is not limited to the case of generating data different from the two survey data that is the original data, but the content that incorporates one of the two survey data that is the original data into the other. It shall include the case of generating fusion data in an updated form.
Data fusion processing and fusion data will be described in detail in later sections.

  Furthermore, when each terminal operated by a person in charge of the research company and a customer company terminal are connected to the server 1 so that they can communicate with each other, each terminal (not shown) or customer company terminal is generated by receiving an input operation. When the data distribution request is received, the generated fusion data can be distributed to each terminal or customer company terminal that is the source of the request.

<< About survey data >>
Next, the first survey data related to the audience rating survey and the second survey data related to the questionnaire survey will be described with reference to FIG.
(A) in FIG. 3 shows an example of the contents of the survey result of the audience rating survey which is the first survey. (B) in FIG. 3 shows an example of the content of the survey result for the questionnaire survey that is the second survey.

  The first survey data D1 obtained by compiling audience rating data indicating the results of the audience rating survey includes information on monitor attributes such as gender, that is, information on demographics (demographic attributes) and television X viewing information. Includes information about presence and absence. Each item is binarized with a numerical value of “0” or “1”. However, the present invention is not limited to this, and each item may be input by a method other than the method of inputting either “0” or “1”.

On the other hand, the second survey data D2 obtained by collecting the individual response data indicating the results of the questionnaire survey is single source data as described above, and the individual response data for several hundred questions is, for example, more than 10,000 people. It is collected and collected from the monitors.
For example, in addition to information on monitor attributes such as gender, that is, information on demographics (demographic attributes) and information on whether or not to watch TV X, psychographic (psychological) Information). Also, in the second survey data D2, each item is binarized with a numerical value of “0” or “1”, but the present invention is not limited to this, and each item is either “0” or “1”. It is good also as inputting by methods other than the method of inputting.

As shown in FIGS. 3A and 3B, the first survey data D1 and the second survey data D2 include items common to both surveys (hereinafter, common items). This common item is a key item in later data fusion processing, and in the case shown in FIG. 3, there are items related to monitor attributes (gender, etc.) and items related to TV program viewing status (whether TV X is viewed, etc.). This is a common item.
On the other hand, the second survey data D2 includes unique items (unique items) that are not included in the first survey and are included only in the second survey. This item is a questionnaire conducted by a research company to investigate the three aspects of consumer attributes, product involvement, and media contact on the same monitor in order to grasp trends in the world, and to grasp consumers from various perspectives. This item is specially set in the survey. For example, the item asking whether or not beer A is purchased shown in FIG. 3B corresponds to this unique item.

<< About server configuration >>
Next, the configuration of the server 1 will be described with reference to FIG. FIG. 4 is a diagram illustrating a hardware configuration of the server 1.
As shown in FIG. 4, the server 1 includes a CPU 1a, ROM 1b, RAM 1c, communication interface (indicated as communication I / F in FIG. 4) 1d, hard disk drive (indicated as HDD in FIG. 4) 1e, mouse, An input device 1f such as a keyboard and an output device 1g such as a display or a printer are included as components. The server 1 is preinstalled with a program for exhibiting the function (hereinafter, data processing program). By reading and executing this data processing program by the CPU 1a, a data processing service by the server 1 is provided.

The hardware configuration of the server 1 is as described above. Hereinafter, the configuration of the server 1 will be described again from the functional aspect with reference to FIG. FIG. 5 is a diagram showing the configuration of the server 1 in terms of functions.
As illustrated in FIG. 5, the server 1 includes a data reception unit 11, a data aggregation unit 12, a data storage unit 13, a process execution unit 14, and a data distribution unit 15. These are responsible for various processes executed by the server 1, and are configured by the cooperation of the above-described hardware components constituting the server 1 and the above-described data processing program. Hereinafter, each of the functional units of the server 1 described above will be described.

  The data receiving unit 11 communicates with a device connected to the server 1 via the information communication networks N1 and N2 and receives data from the device. For example, the data receiving unit 11 is viewed from a measuring device 21 arranged in each household. The rate data is received, and the individual response data is received from the response terminal 31.

  When the data receiving unit 11 receives the audience rating data or the individual response data, the data aggregating unit 12 analyzes the audience rating data or the individual response data, specifies information indicated by the data, and further displays the information in FIG. Are summarized in the table format shown in FIG. That is, the data aggregating unit 12 aggregates the audience rating data transmitted from the measuring device 21 arranged in each household as the first survey data D1 for all monitors and for all items. Further, the individual response data transmitted from the response terminal 31 is collected as the second survey data D2 for all the monitors and for all items.

  The data storage unit 13 stores various data, and includes a hard disk drive 1e mounted on the server 1 as a main component. The data stored in the data storage unit 13 includes the first survey data D1 and the second survey data D2 generated by the data aggregation unit 12. Furthermore, the data storage unit 13 stores fusion data generated by a process execution unit 14 (to be described later) executing a data fusion process.

  In the present embodiment, the first survey data D1 and the second survey data D2 are stored in the hard disk drive 1e in the server 1, but the present invention is not limited to this. That is, the storage device that stores the first survey data D1 and the second survey data D2 may be provided separately from the server 1. For example, a database server that can communicate with the server 1 is used as the first survey data D1. And it is good also as using as a memory | storage device of 2nd investigation data D2.

The process execution unit 14 reads the first survey data D1 and the second survey data D2 stored in the data storage unit 13, executes a data fusion process for fusing these data, and generates fused data.
The data fusion process will be outlined below with reference to FIG. FIG. 6 is a diagram conceptually showing the contents of the data fusion processing.

  The data fusion process is a process of fusing survey data indicating the results of each of the first survey and the second survey performed on different monitors, using the answer contents for each common item as a key. As a result of such processing, as shown in the lower part of FIG. 6, data indicating information obtained by adding the response content to the original item (corresponding to the second item) included in only the second item with respect to the result of the first item. That is, fusion data is generated.

In the data fusion processing, first, the first survey data D1 indicating the survey result of the audience rating survey is compared with the second survey data D2 indicating the survey result of the questionnaire survey. More specifically, a common item included in both surveys is specified between the monitors that have answered the two surveys, and a similarity degree of responses to the common items is calculated.
Here, as described above, the common items include items related to monitor attributes such as sex and items related to the viewing status of television programs (television viewing behavior).

  More specifically, in the data fusion process, for each monitor A (hereinafter referred to as monitor A) that requested the audience survey as the first survey, monitor B (hereinafter referred to as monitor B) that requested the questionnaire survey as the second survey. ) Are searched for the most similar response contents for common items. For this purpose, the degree of similarity between the monitor A and the monitor B is calculated. A method for calculating the degree of similarity will be described in a later section.

  After calculating the degree of similarity, a combination of monitors having a high degree of similarity is searched. More specifically, for the monitor A, the monitor having the highest similarity is searched from the monitors B. The searched monitor (monitor B) and the reference monitor (monitor A) are hereinafter treated as a set of monitors.

Then, the same answer contents as the answer contents of the unique item answered by the monitor B are assigned to the monitor A in the set of monitors.
Here, as described above, the unique items include items related to products / services such as whether or not beer A is purchased, as well as items related to daily life awareness / behavior, media / advertisement involvement, media contact, etc. .

  Monitor A responds to items included in the audience rating survey, specifically, only common items, and does not respond to unique items included only in the questionnaire survey. Therefore, by assigning the same answer contents as the answer contents of the original item answered by the monitor B, a virtual answer to the original item is given to the monitor (monitor A) not answering the questionnaire survey. Become.

By assigning virtual answers to all the monitors A according to the above procedure, the first survey data D1 that aggregates audience rating data indicating the results of audience rating surveys and individual response data that indicates the results of questionnaire surveys are aggregated. The second survey data D2 is merged.
As a result, fusion data is generated as pseudo survey data including common items and original items in the survey results of different surveys performed on different monitors, and the generated fusion data is stored in the data storage unit 13. The

  It should be noted that the number of fused data is not necessarily the same as the number of monitors A or the number of monitors B. For example, the number of monitors B is a multiple of the number of monitors A. If not, fusion data including a number of data corresponding to the least common multiple of the number of monitors may be generated.

  The data distribution unit 15 receives from the data storage unit 13 when the data reception unit 11 receives a fusion data distribution request issued from each terminal (not shown) or customer company terminal that is communicably connected to the server 1. The merged data is read out and distributed to each of the above terminals and customer company terminals.

<< Survey data processing method >>
Next, a survey data processing method according to this embodiment will be described.
The survey data processing method according to the present embodiment is performed using the server 1 that is a computer. In other words, the survey data processing method according to the present embodiment is applied to data processing (hereinafter referred to as data processing service) executed by the server 1 as an ASP service. Hereinafter, as a description of the survey data processing method according to the present embodiment, the flow of the data processing service by the server 1 and each process in the service will be described.

  The data processing service by the server 1 proceeds according to the flow shown in FIG. FIG. 7 is a diagram showing a rough flow of a series of processes related to the data processing service.

First, the data processing service starts from a step of acquiring first survey data (S001). Here, the first survey data is obtained by tabulating audience rating data indicating the results of the audience rating survey.
If this process S001 is demonstrated in detail, the data reception part 11 will receive audience rating data from the measuring machine 21 arrange | positioned in each household. Then, after a certain period of time, the data aggregating unit 12 aggregates these audience rating data and generates first survey data for all monitors. The first survey data acquired in this step S001 is stored in the data storage unit 13.
Thus, in the present embodiment, the first survey data is acquired by the cooperation of the data receiving unit 11 and the data aggregating unit 12. From this viewpoint, the data receiving unit 11 and the data aggregating unit 12 constitute a first survey data acquisition unit that acquires first survey data.

After acquiring the first survey data, the step of acquiring the second survey data (S002) is executed.
If this process S002 is demonstrated in detail, the data reception part 11 will receive separate response data from the response terminal 31 of each monitor. Thereafter, when the individual response data for all items are collected from all the monitors, the data aggregating unit 12 aggregates the individual response data and generates second survey data for all the monitors and for all the items. Then, the second survey data acquired in this step S002 is stored in the data storage unit 13.
Thus, in the present embodiment, the second survey data is acquired by the cooperation of the data reception unit 11 and the data aggregation unit 12. From this viewpoint, the data receiving unit 11 and the data aggregating unit 12 constitute a second survey data acquiring unit that acquires the second survey data.

In this embodiment, the second survey data is acquired after acquiring the first survey data. However, the present invention is not limited to this, and the first survey data is acquired after acquiring the second survey data. The first survey data and the second survey data may be acquired at the same time.
Further, when one or both of the first survey data and the second survey data is dynamically changing data, only the changed data may be acquired.

When two pieces of survey data (first survey data and second survey data) are stored in the data storage unit 13 through the above-described steps S001 and S002, the processing execution unit 14 of the server 1 thereafter performs these first surveys. The data and the second survey data are read and the data fusion process is executed (S003).
In this step S003, fusion data indicating the contents of responses in both the audience rating survey and the questionnaire survey is generated, and the fusion data generated in this step S003 is stored in the data storage unit 13. This step S003 may be automatically executed when the first survey data and the second survey data are stored in the data storage unit 13 by the previous steps S001 and S002, or the server 1 It may be executed when a predetermined input operation by a user (for example, an employee of a research company) is received on the server 1 side.

After that, for example, when distribution data distribution is requested through each terminal or customer company terminal as necessary, the data distribution unit 15 of the server 1 receives the request from the data reception unit 11 and the data storage unit 15 The fusion data is read from 13 and the data is distributed to each terminal or customer company terminal (S004).
At the stage where the above series of steps is completed, one data processing service at a certain time (for example, time t) is completed.

Next, step S003 of executing the data fusion process in the data processing service described above will be described in more detail with reference to FIG. FIG. 8 is a diagram showing a flow of data fusion processing at a certain time point (time point t).
The data fusion process executed by the process execution unit 14 of the server 1 proceeds according to the flow of FIG. Hereinafter, each step will be described.

In the data fusion process, first, the first survey data and the second survey data stored in the data storage unit 13 are read, and common items included in both the first survey data and the second survey data are designated (S101).
Specifically, the process execution unit 14 analyzes the content of the first survey data, and specifies the response content that one monitor A included in the first survey data has answered to the common question. Similarly, the second survey data is analyzed, and the content of the answer that one monitor B included in the second survey data has answered to the common question is specified.
In the present embodiment, monitor attributes such as gender and the content of responses regarding TV viewing behavior are designated as common items.

Next, in order to identify which person on monitor A is most similar to which person on monitor B, a composite variable (principal component obtained by principal component analysis) is used using items common to both monitor A and monitor B. A score calculation function or the like) is created (S102), and the composite variable (principal component score calculation function or the like obtained by principal component analysis) calculated in step S102 or an arbitrary function set in advance is stored in the data storage unit 13 (S103). ).
Specifically, the fusion parameter used when the process execution unit 14 calculates a value (score group for distance calculation) necessary for obtaining the distance between the monitor A of the first survey and the monitor B of the second survey. Set the function to be For the creation of a composite variable, for example, a principal component score generated by performing principal component analysis in statistics may be used, or an arbitrary coefficient may be multiplied to each common item, and then the total obtained may be used. . Moreover, the composite variable may be single or plural.

Next, a distance calculation score group is calculated by applying the above function to the common items of both the first survey data and the second survey data (S104).
Specifically, the process execution unit 14 obtains the values of the monitors A and B of the first survey data and the second survey data by using the function and the common item. At this time, the distance calculation score group becomes an objective variable, and the common item becomes an explanatory variable.

Next, the distance calculation score groups calculated in step S104 are compared, and distance calculation is performed for distances indicating the degree of similarity between each monitor A of the first survey data and each monitor B of the second survey data ( S105).
In the present embodiment, the difference in answer contents for each common item is expressed as a distance, and the total distance obtained by summing up the distance calculation score groups for each common item is used as the degree of similarity between monitors, and the total distance is a small value. This indicates that the monitor has a close attribute. Specifically, the process execution unit 14 adds up the distance calculation score groups of the respective items, and uses the result as the degree of similarity between the monitors.

Next, after performing the distance calculation in the above-described step S105, for each monitor A of the first survey data, the monitor having the shortest total distance is specified from the monitors B of the second survey data, and is regarded as the same monitor. Then, they are fused (S106).
At this time, in the present embodiment, when setting the allocation pattern between the monitor A of the first survey data and the monitor B of the second survey data, “unconstrained statistical matching” or “constrained statistical matching” Any method may be used.

For example, it is possible to set an allocation pattern by adopting a method of “unconstrained statistical matching” that can simplify processing and efficiently process a large amount of data. Here, in order to simplify the case, the case shown in FIG. 9 will be described as an example.
FIG. 9 is an explanatory diagram when the data fusion processing is performed by adopting the “unconstrained statistical matching” as described above.

As shown in FIG. 9, the process execution unit 14 monitors the second survey data monitor B having the highest degree of similarity calculated in the above-described step 106 for the monitor A of the first survey data, that is, the most similar second survey. Select and link one data monitor B. For example, for the first survey data monitor A (A00001), the second survey data monitor B (B00002) is used. For the first survey data monitor A (A00002), the second survey data monitor B is used. Identify (B00003) as the most similar monitor and merge.
As a result, the same answer content as the answer of the second survey data monitor B (B00002) is assigned to the first survey data monitor A (A00001) as a virtual answer, and the first survey data monitor A (A00002) is assigned. The same answer as the answer of monitor B (B00003) of the second survey data is assigned as a virtual answer. That is, as a result of allocating virtual answers to the monitors A00001 and A00001, the data are merged, and finally, merged data shown in the lower part of FIG. 9 is generated.

  On the other hand, the fusion data can be set by assigning an allocation pattern using the “constrained statistical matching” method so that the average and variance of the original data (first survey data and second survey data) can be maintained. it can. In this “constrained statistical matching”, the process execution unit 14 sets an allocation pattern according to a statistical solution based on the degree of similarity between monitors.

Specifically, the process execution unit 14 adopts a transportation problem solution as a method for setting an assignment pattern, and assigns an assignment pattern that satisfies both the following preconditions (A) and (B) by the solution. I am going to set it.
(A) About the total result regarding a 1st investigation, it is set as the same between the 1st survey data before fusion, and fusion data.
(B) About the total result regarding a 2nd investigation, it is made the same between the 2nd investigation data before fusion, and fusion data.
Here, the total result regarding the first survey or the second survey is the ratio of the number of persons when the number of monitors is totaled according to the content of the response to the survey.
In setting the allocation pattern by solving the transportation problem, a weight is set for each monitor in the first survey and each monitor in the second survey. Here, the weight set for each monitor A of the first survey data corresponds to the demand amount in the transportation problem, and the weight set for each monitor B of the second survey data corresponds to the supply amount. To do.

Here, in order to simplify the case, the case shown in FIG. 10 will be described as an example.
FIG. 10 is an explanatory diagram when the data fusion processing is performed by adopting the “constrained statistical matching” as described above.

For example, as shown in FIG. 10, one of the two survey data related to data fusion (first survey data) is data of two monitors (A00001 to A00002), and the other (second survey data) is When there are three monitors (B00001 to B00003), the monitor A of the first survey data is linked to one or more of the monitors B of the second survey data, and the monitor B of the second survey data is monitored as the monitor of the first survey data Connect with one or more of A. At this time, the monitor B of the second survey data may be connected to one monitor of the first survey data with a weight of 3, and as shown in FIG. 10, one monitor of the first survey data and the weight of 1 (A00001). And B00002, A00002 and B00002), and another monitor of the first survey data may be connected by weight 2 (A00001 and B00001, A00002 and B00003). Further, the monitor A of the first survey data may be connected to one monitor of the second survey data by weight 2 (A00001 and B00001, A00002 and B00003), or one monitor of the first survey data and weight 1 (A00001). And B00002) and one other monitor of the first survey may be connected with weight 1 (A00002 and B00002).
However, the combined weights are combined so that the sum of the similarities (similarities calculated from the above values) between the monitor A of the first survey data and the monitor B of the second survey data is minimized. As a result, the fusion data can maintain the average and variance of the original data (first survey data and second survey data). In this “constrained statistical matching”, the weight distribution set in this way corresponds to the allocation pattern.

  As a result, the first survey data monitor A (A00001) is assigned the same answer content as the second survey data monitor B (B00001) response and the second survey data monitor B (B00002) response as a virtual answer. The former is treated as an answer for two monitors, and the latter is treated as an answer for one monitor. The first survey data monitor A (A00002) is assigned the same answer contents as the responses of the second survey data monitor B (B00002) and the second survey data monitor B (B00003) as virtual answers. The former is treated as an answer for one monitor, and the latter is treated as an answer for two monitors. That is, as a result of the virtual answers being assigned to the monitors A00001 and A00001, the data are merged, and finally, the merged data shown in the lower part of FIG. 10 is generated.

  Thereafter, a process of saving the fusion result (fusion data) (S107) is executed, and the fusion data is stored in the data storage unit 13. At this point, the data fusion process is completed.

  Note that the fusion processing may be performed according to the value of a specific item such as gender among the common items. In that case, the process of said process S104-S106 is performed according to the value. In this case, the fusion result is saved as one file.

Here, in the present embodiment, the first survey and the second survey are surveys that rotate the monitor, and the first survey data D1 and the second survey data D2 are data that dynamically change over time. It becomes.
Specifically, if the same monitor is surveyed for a long time, it may be affected by survey familiarity and learning effects, and the sample monitor may not respond to changes in the population. In order to avoid such a situation, in the present embodiment, after a certain period of time, some or all of the monitors A and B of the first and second surveys are sequentially rearranged (rotation). In addition to the periodic rotation as described above, there is a case where the monitor is irregularly replaced by adding a new monitor in order to supplement the dropped monitor.
Note that the dynamically changing data in this embodiment is not limited to survey data such as questionnaire surveys, operation log data that can be automatically acquired from various devices, Web logs, purchase history, application usage history, or the like. A large amount of customer data that changes constantly (so-called big data) is also included and can be applied in this embodiment.

In the present embodiment, the timing of change is different between the first survey and the second survey. Specifically, in the first survey, a part of the monitor A is sequentially replaced at a predetermined rate every time 1 elapses from the time point t. In the second investigation, every time n elapses from time t, the entire monitor B is totally replaced.
In this embodiment, the description will be made assuming that all the monitors B are totally replaced. However, the present invention is not limited to this, and the monitor B of the second survey has the same timing as the monitor A of the first survey. It may be different, and for example, half or 1/3 of the monitor B may be partially replaced every time n elapses from time t.

Hereinafter, specific examples in the case where the first survey data is dynamically changing data will be described with reference to FIGS. 11 to 13 as examples.
11 and 13 are diagrams conceptually showing a state in which the first survey data has changed. FIG. 12 is a diagram showing a flow of data fusion processing at time t + 1 after the fusion is performed.

  As shown in FIG. 11, at the time of performing the fusion (time t), the monitor A of the first survey and the monitor B of the second survey are all merged, but after the fusion is performed (time t + 1), they are merged. A part of the monitor A which has been lost does not exist in the first survey data, and a new monitor A ′ which does not exist at the time of performing the fusion (time t) is added to the first survey data. . Therefore, it is necessary to specify a partner to be associated with the new monitor A ′ from the second survey data.

In the above case, step S003 of executing the data fusion process among the data processing services described above at the time point t + 1 after the fusion will be described with reference to FIG.
At the time point t + 1 after performing the fusion, the data fusion process executed by the process execution unit 14 of the server 1 proceeds according to the flow of FIG.

In the data fusion process at time t + 1 after the fusion is performed, first, new first survey data is specified from the first survey data (S201).
Specifically, the process execution unit 14 analyzes the content of the first survey data stored in the data storage unit 13, and creates new first survey data (the response content of the monitor A ′) from the first survey data. Data).

Next, the distance calculation score group is calculated by applying the function stored in S103 of the data fusion process at the time point t to the common items of both the new first survey data and the second survey data. (S202).
Specifically, the process execution unit 14 obtains a composite variable for each of the monitors A ′ and B of the new first survey data and the second survey data by using the function and the common item. At this time, the process similar to S104 of the data fusion process at the time point t described above is executed except that the process is performed only for the new first survey data.

  Thus, even when a new monitor A ′ is added to the first survey data, the function used when calculating the distance calculation score group is not recalculated, and the function that has already been used and proven effective. By reusing, the processing is simplified, and even a large amount of new survey data can be processed quickly.

At this time, the new first survey data is exactly the same as the fusion data (or the second survey data) because, for example, the common item depends on the time point t as in the TV viewing tendency in the specific period A. There may be cases where common items are not held. That is, as shown in FIG. 11, at the time of performing the fusion (time t), the common item r (t) is included in both the first survey data (monitor A) and the second survey data (monitor B). However, after the fusion is performed (time t + 1), the common item of the first survey data of the new monitor A ′ changes to the common item r (t + 1) as time elapses, and the fusion data (or second data) Survey data), the same common items are not held.
Therefore, in the present embodiment, at a time point after time t (after t + 1), for example, a statistic such as an average value of the TV viewing tendency in the specific period B is calculated and substituted for the common item r (t) at time t. To do. Strictly speaking, this is different from the TV viewing tendency in the specific period A at the time point t, but it can be replaced by a viewing tendency at an approximate time, which is not different from the case where the same common item is held, and is new. The survey data can be merged.

Next, the distance calculation score groups calculated in the above step S202 are compared, and distance calculation is performed for the distance indicating the degree of similarity between each monitor A ′ of the new first survey data and each monitor B of the second survey data. Execute (S203).
Specifically, the process execution unit 14 adds the distance calculation score groups of the respective items for each monitor A ′ of the new first survey data, and uses the result as the degree of similarity between the monitors.

Next, after performing the distance calculation in the above-described step S203, for each monitor A ′ of the new first survey data, a monitor having a short total distance is identified from the monitors B of the second survey data, and the same monitor They are considered to be tied and fused (S204).
At this time, in the present embodiment, when setting the allocation pattern between the new monitor A ′ for the first survey data and the monitor B for the second survey data, for the purpose of simplifying the processing, "Matching" is used to implement data fusion processing.
Specifically, the process execution unit 14 monitors the second survey data monitor B having the highest degree of similarity calculated in the above step 203 with respect to the new first survey data monitor A ′, that is, the most similar second survey data A ′. One monitor B of survey data is selected and linked, and that monitor B is identified as the most similar monitor and merged. As a result, the same response content as the response of any one of the monitors B in the second survey data is assigned to the new monitor A ′ of the first survey data as a virtual response, and finally new fusion data is generated. Is done.

  In the present embodiment, when the monitor A of the first survey data changes to the monitor A ′ at the time point t + 1, the monitors A and B that are the fusion partners still exist at that time as they are. The data fusion process at time t + 1 is executed again only for the monitor A ′ of the new first survey data. As a result, the fusion data that has already undergone the fusion processing can be utilized, and the new monitor A ′ of the first survey data can be fused accurately and efficiently.

  Thereafter, a process of saving the fusion result (fusion data) (S205) is executed, and the fusion data is stored in the data storage unit 13. Thereby, the data fusion process at the time point t + 1 is completed.

As another example, a specific example in the case where the second survey data is dynamically changing data will be described below with reference to FIG.
FIG. 14 is a diagram conceptually illustrating a state in which the second survey data has changed.

As described above, the first survey and the second survey have different timings of change, and the second survey totally replaces all the monitors B every time n elapses from time t.
As shown in FIG. 11 described above, the first survey monitor A and the second survey monitor B are all merged at the fusion execution time point (time point t). However, as shown in FIG. At (time t + n), all of the merged monitors B no longer exist in the second survey data, and the second survey data is a new monitor B ′ that did not exist at the merge execution time (time t). Total replacement. Therefore, for the new monitor B ′, it is necessary to specify the partner to be associated from the first survey data.
In this case, in principle, a process similar to the data fusion process executed by the process execution unit 14 of the server 1 at the fusion execution time t shown in FIG. 8 is newly executed. Note that the same process as the data fusion process executed by the process execution unit 14 of the server 1 at the time point t + 1 after the fusion shown in FIG. 12 may be executed for the monitor B ′ of the second survey data.

  Further, after the fusion is performed (time t + n + 1), a part of the monitor A that has been merged with the monitor B ′ of the second survey data no longer exists in the first survey data, and the fusion execution time (time t + n). ), A new monitor A ′ that did not exist is added to the first survey data. Therefore, for the new monitor A ′, it is necessary to specify the partner to be linked from the second survey data. In this case as well, the processing execution unit 14 of the server 1 performs the processing at the time t + 1 after performing the fusion shown in FIG. A process similar to the data fusion process to be executed is executed.

<< Other Embodiments >>
In the above embodiment, the investigation data processing apparatus and the investigation data processing method of the present invention have been mainly described. However, the above embodiment is for facilitating the understanding of the present invention, and does not limit the present invention. The present invention can be changed and improved without departing from the gist thereof, and the present invention includes the equivalents thereof.

  In the above-described embodiment, an example in which the first survey company and the second survey are performed by one survey company has been described. However, the present invention is not limited to this. For example, the first survey is conducted by a survey company and the second survey is conducted by another company, and the present invention can be applied even when survey data indicating survey results conducted by different entities are merged. It is. In this way, for client companies, the results of questionnaire surveys conducted by the company (for example, questionnaire surveys conducted for the purpose of grasping the public reaction regarding products sold and services provided by the company) and the survey company Can be integrated with surveys conducted by the company (for example, audience rating surveys), and effective media planning can be made in accordance with the needs of each client company.

  In the above embodiment, the unique item exists only in the second survey, and the first survey includes only the common item and does not include the unique item. However, the present invention is not limited to this. In other words, the present invention is applicable even in the case where there is a unique item (first survey unique item) that is not included in the second survey and included only in the first survey. Is possible.

  In the above embodiment, the fusion data is generated by fusing two pieces of survey data. However, the present invention is not limited to this. That is, even if there are three or more survey data, if there are common items, they can be merged to generate fused data, so the present invention can be applied to such a case. Is possible.

  In the above-described embodiment, the case of generating the fusion data as the pseudo single source data by fusing the single source data and the non-single source data has been described as an example. However, the present invention is not limited to this. . That is, the present invention can be applied to both a case where non-single source data is fused to generate fused data and a case where single source data is fused to generate fused data.

  In the above embodiment, in evaluating the degree of similarity between the monitor A of the first survey and the monitor B of the second survey, the response contents of all common items included in both surveys are evaluated. However, the present invention is not limited to this. That is, only a part of a plurality of common items may be evaluated.

1 server (survey data processing device)
1A CPU
1B ROM
1c RAM
1d Communication Interface 1e Hard Disk Drive 1f Input Device 1g Output Device 11 Data Receiving Unit 12 Data Aggregation Unit 13 Data Storage Unit 14 Processing Execution Unit 15 Data Distribution Unit 21 Measuring Device 31 Answer Terminals D1, D2 Survey Data N1, N2 Information Communication network

Claims (7)

A survey data processing device that processes survey data indicating the results of each of the first survey and the second survey conducted on different monitors,
A first survey data acquisition unit that acquires first survey data collected for the number of persons of the first survey monitor who answered the first survey for the contents of responses to common items included in both the first survey and the second survey When,
A second survey data is acquired by collecting the contents of the responses to the common items and the second survey unique items included only in the second survey for the number of second survey monitors who answered the second survey. 2 Survey data acquisition unit,
The first survey data and the second survey data obtained by merging the first survey data acquired by the first survey data acquisition unit and the second survey data acquired by the second survey data acquisition unit. A process execution unit for executing a data fusion process for generating fusion data indicating the content of each of the answers,
The data fusion process includes:
A value for the content of the response to the common item is calculated using a predetermined arithmetic expression, and the result of comparing the values between the first survey monitor and the second survey monitor is used to answer the common item A calculation process for calculating the degree of similarity of content;
An assignment process for assigning, to the first survey monitor, the same response content as the response content to the second survey original item of the second survey monitor in the allocation pattern set based on the calculated similarity degree; Including,
When both or one of the first survey monitor and the second survey monitor changes, the first survey monitor and the second survey that have already been allocated by the allocation process and the allocation partner still exists at that time. An investigation data processing apparatus characterized in that an assignment partner is directly taken over for a monitor, and the calculation process and the assignment process are executed only for a new first investigation monitor or a new second investigation monitor. Before and after both or one of the first survey monitor and the second survey monitor change,
The survey data processing apparatus according to claim 1, wherein the predetermined arithmetic expression used in the calculation process does not change. Before and after either or both of the first survey monitor and the second survey monitor change, when the survey time for the common item is different,
In the calculation process and the allocation process, the process execution unit substitutes an answer content for an item or time approximate to the common item or a statistical value of the answer content as an answer content for the common item. Item 3. A survey data processing device according to item 1 or 2. Before both or one of the first survey monitor and the second survey monitor change,
The process execution unit, in the allocation process, the first survey totalization result is aligned between the fusion data and the first survey data, and the aggregation results the second investigation concerning the second survey The survey data processing apparatus according to any one of claims 1 to 3, wherein the allocation pattern that matches between the data and the fusion data is set according to a statistical solution.   5. The survey data processing apparatus according to claim 1, wherein the first survey data includes a response content for a first survey unique item included only in the first survey. 6. The first survey is a survey on media contact,
The survey data processing apparatus according to claim 1, wherein the second survey is a questionnaire survey that captures multifaceted aspects of consumer attributes, product involvement, and media contact. A survey data processing method for processing survey data indicating a survey result of each of the first survey and the second survey performed on different monitors by a computer,
The computer is
Obtaining the first survey data collected for the number of persons of the first survey monitor who answered the first survey, the content of the response to the common items included in both the first survey and the second survey;
The step of acquiring the second survey data in which the contents of the responses to the common item and the second survey unique item included only in the second survey are collected by the number of the second survey monitors who answered the second survey. When,
Fusing the acquired first survey data and the second survey data, and executing a data fusion process for generating fused data indicating the contents of each of the first survey data and the second survey data; Run,
The data fusion process includes:
A value for the content of the response to the common item is calculated using a predetermined arithmetic expression, and the result of comparing the values between the first survey monitor and the second survey monitor is used to answer the common item A calculation process for calculating the degree of similarity of content;
An assignment process for assigning, to the first survey monitor, the same response content as the response content to the second survey original item of the second survey monitor in the allocation pattern set based on the calculated similarity degree; Including,
When both or one of the first survey monitor and the second survey monitor changes, the first survey monitor and the second survey that have already been allocated by the allocation process and the allocation partner still exists at that time. An investigation data processing method characterized in that the assignment partner is directly taken over for the monitor, and the calculation process and the assignment process are executed only for a new first investigation monitor or a new second investigation monitor.

Images