JP6452025B2 - Display system, display method, and display program - Google Patents

Description

  The present invention relates to a display system, a display method, and a display program.

  As a data display method, there is a method of displaying data by 3D Parallel Coordinates or a heat map.

  In general, in a heat map with an attribute as the X axis, the Y axis represents only one attribute. When the height is the X axis, for example, only one attribute of weight is represented on the Y axis.

  The attributes include a continuous variable attribute and a categorical variable attribute. The attribute value of the continuous variable is a numerical value. The attribute value of the categorical variable is an item. “Weight” in the above example is an attribute of a continuous variable. As an example of the attribute of the categorical variable, for example, “sex” can be cited, and the attribute value is an item of “male” or “female”. The attribute represented by the Y axis of the heat map may be an attribute of a categorical variable such as “gender”. In this case, the items “male” and “female” are represented on the Y axis. For example, when the height is the X axis, the count value of the item “male” and the item “female” is expressed as a heat map for each section obtained by dividing the X axis representing the height.

  FIG. 15 is a schematic diagram illustrating an example of a heat map in which the Y axis represents the attribute of a continuous variable. In the example illustrated in FIG. 15, “time” is the X axis, and “number of customers” is the Y axis. “Number of customers” is an attribute of a continuous variable.

  FIG. 16 is a schematic diagram illustrating an example of a heat map in which the Y-axis represents an attribute of a categorical variable. In the example shown in FIG. 16, “time” is the X axis, and “store code” is the Y axis. “Store code” is an attribute of a categorical variable, and the attribute value is an item such as “Store A”, “Store B”,. In the example shown in FIG. 16, for example, a set of data in which a time zone is associated with an item such as “Store A” is given in advance, and the count value of “Store A” corresponding to each time zone, “ The count value of “Store B” is expressed as a heat map. Also in the example illustrated in FIG. 16, the Y axis represents only one attribute “store code”.

  Patent Document 1 describes that pixel data is classified into a plurality of clusters by clustering processing based on the K-means method.

  Patent Document 2 describes using a heat map as an outlier detection visualization tool.

JP 2009-212605 A (see paragraph 0066) JP 2009-76056 A (see paragraph 0016)

  As described above, generally, in a heat map in which a certain attribute is an X axis, the Y axis represents only one attribute. Accordingly, the observer of the heat map can grasp only the change state of one attribute. For this reason, the observer cannot easily grasp an attribute having a similar change pattern among a plurality of attributes.

  Therefore, an object of the present invention is to provide a display system, a display method, and a display program that enable an observer to easily grasp attributes having a similar change pattern among a plurality of attributes.

The display system according to the present invention is provided with a set of attribute value pairs of a plurality of attributes associated with each other, and when one attribute is specified among the plurality of attributes, the attribute values of attributes not specified are displayed. The aggregated data generating means for generating the aggregated data aggregated according to the attribute value category of the designated attribute for each attribute not designated, the clustering means for clustering the aggregated data, and the designated attribute as the X axis, For each cluster obtained as a result of clustering, display means for displaying the aggregated data as a heat map, selection accepting means for accepting selection of aggregated data, and narrowing down the aggregated data to be displayed based on the selected aggregated data A narrowing-down means, and the narrowing-down means displays a vector represented as the selected aggregated data and other aggregated data. Based on the distances between the vectors, characterized in that narrow down the aggregated data to be displayed.

Further, the display method according to the present invention is provided with a set of attribute value pairs of a plurality of attributes that are associated with each other, and when one attribute among the plurality of attributes is specified, the attributes of the attributes that are not specified Aggregated data that is aggregated according to the attribute value category of the specified attribute is generated for each attribute that is not specified, the aggregated data is clustered, and the specified attribute is used as the X axis. Displays aggregate data as a heat map for each cluster , accepts selection of aggregate data, selects aggregate data to be displayed based on the selected aggregate data, and selects when narrowing the aggregate data to be displayed The aggregate data to be displayed is narrowed down based on the distance between the vector represented as the aggregated data and each vector represented as other aggregated data. Characterized in that the writing.

Further, the display program according to the present invention is not specified when a set of attribute values of a plurality of attributes associated with each other is given to the computer and one attribute is specified among the plurality of attributes. Aggregated data generation process that generates aggregated data for each attribute that is not specified, aggregated data that is aggregated according to the attribute value category of the specified attribute , clustering process that clusters the aggregated data , and specified attribute X For each cluster obtained as a result of clustering, display processing that displays aggregated data as a heat map, processing that accepts selection of aggregated data, and aggregated data to be displayed based on the selected aggregated data It is displayed as the selected aggregate data in the process of narrowing down the aggregate data to be displayed by executing the narrowing process. And vector, based on the distance between the vector represented as other aggregate data, and characterized in that narrowed aggregated data to be displayed.

  According to the present invention, an observer can easily grasp an attribute having a similar change pattern among a plurality of attributes.

It is a block diagram which shows the structural example of the display system of the 1st Embodiment of this invention. It is a schematic diagram which shows the example of input data. It is a flowchart which shows the example of process progress of the display system of 1st Embodiment. It is a schematic diagram which shows the example of the display screen of the display system in 1st Embodiment. It is a schematic diagram which shows the total data of other each attribute (Each attribute such as “Precipitation” illustrated in FIG. 2) when “Date” is designated in Step S1. It is a schematic diagram which shows the state after performing the process of step S3 with respect to the total data typically shown in FIG. It is a schematic diagram which shows the result of having performed the clustering process of step S4 with respect to the total data typically shown in FIG. It is a schematic diagram which shows the example of the heat map displayed by step S5. It is a figure which shows the example of the heat map of the total data displayed for every cluster. It is a block diagram which shows the structural example of the display system of the 2nd Embodiment of this invention. It is a schematic diagram which shows the example of the display screen of the display system in 2nd Embodiment. It is a schematic diagram which shows the example of a heat map display at the time of narrowing down the total data made into display object after displaying the heat map illustrated in FIG. It is a schematic block diagram which shows the structural example of the computer which concerns on each embodiment of this invention. It is a block diagram which shows the example of the minimum structure of the display system of this invention. It is a schematic diagram which shows the example of the heat map in which the Y-axis represents the attribute of a continuous type variable. It is a schematic diagram which shows the example of the heat map in which a Y-axis represents the attribute of a categorical variable.

  Hereinafter, embodiments of the present invention will be described with reference to the drawings. Hereinafter, a person who operates the display system of the present invention is referred to as an operator.

Embodiment 1. FIG.
FIG. 1 is a block diagram showing a configuration example of a display system according to the first embodiment of the present invention. The display system 1 of the present embodiment includes a designation receiving unit 2, a total data generation unit 3, a decomposition unit 4, a clustering unit 5, and a display unit 6.

  A set of attribute value pairs of a plurality of attributes associated with each other is input to the display system 1 as input data. FIG. 2 is a schematic diagram illustrating an example of input data. The input data can be expressed as a table as shown in FIG. In the example shown in FIG. 2, “date”, “precipitation”, “temperature”, “sales promotion campaign”, “number of customers”, and “wind speed” are illustrated as attributes. It is not limited to the example shown. In the example shown in FIG. 2, the date is shown as “date”, but “year”, “month”, and “day” may be independent attributes. In each row shown in FIG. 2, the attribute values of “date”, “precipitation”, “temperature”, “sales promotion campaign”, “number of customers”, and “wind speed” are associated with each other. That is, each row shown in FIG. 2 represents a set of attribute values. The row number shown in FIG. 2 is an ID of a set of attribute values. Each attribute such as “date” corresponds to a column of the table. Among the attributes exemplified in FIG. 2, “sales promotion campaign” is an attribute of a categorical variable, and the other attributes are attributes of a continuous variable. In the following explanation, the items “Festival”, “TV Commercial”, “None”, “Bonus Distribution”, “Coupon Distribution”, “Magazine Advertising”, and “Newspaper Advertising” are attribute values of “Sales Promotion Campaign”. It is assumed that Note that the input data may include only the attributes of continuous variables or only the attributes of categorical variables.

  The designation accepting unit 2 accepts designation of one attribute among a plurality of attributes included in the input data from an operator via an input device (not shown in FIG. 1). The attribute specified by the operator is referred to as a specified attribute. The designation receiving means 2 determines the designated attribute as an attribute corresponding to the X axis of the heat map.

  When the designation receiving unit 2 receives the designation of the attribute from the operator, the total data generating unit 3 generates total data obtained by totaling attribute values of attributes other than the designated attribute according to the attribute value classification of the designated attribute. The total data generation unit 3 generates total data for each attribute other than the designated attribute. The classification when designated as a designated attribute is predetermined for each attribute. In the case of attributes of continuous variables, the individual divisions may be determined by dividing possible attribute values at predetermined intervals. For example, in the case of “date”, attribute values are delimited every month, such as “June 1, 2014 to June 30, 2014”, “July 1, 2014 to July 31, 2014”, and the like. Thus, the section may be determined in advance. Further, for example, in the case of “temperature”, a section may be determined in advance by dividing attribute values at regular intervals, such as “0 ° C. to 5 ° C.” and “5 ° C. to 10 ° C.”. In the case of a categorical variable, each attribute value may be determined as one category.

  When the aggregate data generation means 3 generates aggregate data regarding the attribute of the continuous variable, the aggregate data is calculated by calculating the statistical value of the attribute value corresponding to the category as the aggregate value for each attribute value category of the designated attribute. Is generated. Regarding the attributes of continuous variables, calculating the statistical value of the attribute value corresponding to the category corresponds to the aggregation of attribute values according to the category. Hereinafter, the case where this statistical value is an average value will be described as an example.

  For example, assume that the designated attribute is “date”. When the aggregated data generation means 3 generates aggregated data of precipitation, which is a continuous variable, the input data illustrated in FIG. 2 is referred to, and the classification “June 1, 2014 to June 30, 2014” For the other categories such as “July 1, 2014 to July 31, 2014”, the average value of precipitation corresponding to is calculated for each category. calculate. The total data generation means 3 uses a vector whose element is the average value of precipitation obtained for each category as total data of precipitation.

  Also, for example, assume that the designated attribute is “temperature”. The aggregate data generation means 3 refers to the input data illustrated in FIG. 2 when generating aggregate data of precipitation that is a continuous variable, and calculates the average value of precipitation corresponding to the category of “0 ° C. to 5 ° C.” Is calculated as an aggregate value, and similarly for other categories such as “5 ° C. to 10 ° C.”, an average value of precipitation is calculated for each category. The total data generation means 3 uses a vector whose element is the average value of precipitation obtained for each category as total data of precipitation.

  As can be seen from the above example, the aggregated data of the attributes of continuous variables is represented as a vector of the number of divisions. The total data generation means 3 uses a vector in which the average values are arranged in the order of classification as the total data.

  In the above two examples, the case of generating the precipitation total data has been described as an example. However, the total data generation unit 3 similarly calculates the total data for the attributes of the continuous variables other than the precipitation. Generate.

  The aggregate data generation means 3 counts the number of items that are attribute values of the categorical variable corresponding to the category for each category of the attribute value of the designated attribute when generating aggregate data regarding the attribute of the categorical variable, By calculating the count value as the total value, the total data is generated. Regarding the attributes of categorical variables, counting the number of items corresponding to a category corresponds to the aggregation of attribute values according to the category.

  For example, assume that the designated attribute is “date”. When generating the total data of the “sales promotion campaign” that is a categorical variable, the total data generation unit 3 refers to the input data illustrated in FIG. 2 and reads “June 1, 2014 to June 30, 2014”. The number of items (attribute values) “festival” corresponding to the category “,” the number of items “TV commercial”, and the like are counted for each item. Similarly, other data such as “July 1, 2014 to July 31, 2014” is counted by the total data generation unit 3 for each category and counts the number of items corresponding to the category. As a result, the same number of count values as the number of sections is obtained for one item, and a vector in which the count values are arranged in the order of the sections is obtained. The number of dimensions of this vector is the same as the number of sections. The total data generation means 3 uses a set of vectors for each item as total data for the “sales promotion campaign”.

  Also, for example, assume that the designated attribute is “temperature”. When generating the total data of the “sales promotion campaign” which is a categorical variable, the total data generation means 3 refers to the input data illustrated in FIG. 2 and corresponds to the “Festival” corresponding to the category “0 ° C. to 5 ° C.”. The number of items (attribute values) ", the number of items" TV commercial ", etc. are counted for each item. Similarly, the total data generation unit 3 counts the number of items corresponding to each category for each category such as “5 ° C. to 10 ° C.”. Similar to the above example, the count value equal to the number of sections is obtained for one item, and a vector in which the count values are arranged in the order of the sections is obtained. The number of dimensions of this vector is the same as the number of sections. The total data generation means 3 uses a set of vectors for each item as total data for the “sales promotion campaign”.

  In the above two examples, the case where the aggregated data of “sales promotion campaign” is generated has been described as an example. However, the aggregated data generation unit 3 similarly generates aggregated data for the attributes of other categorical variables. To do.

  The decomposing means 4 classifies the aggregate data generated for each attribute into continuous variable aggregate data and categorical variable aggregate data. The disassembling means 4 classifies the aggregated data represented as one vector as aggregated data of continuous variables, and classifies the aggregated data represented as a set of a plurality of vectors as aggregated data of categorical variables. That's fine. The disassembling unit 4 may add label information indicating whether the data is the aggregated data of the attribute of the continuous variable or the aggregated data of the attribute of the categorical variable to the classified aggregated data. Hereinafter, description will be made assuming that the decomposing means 4 adds such label information to the classified total data.

  Further, the decomposing means 4 decomposes the aggregate data of the attribute of the categorical variable for each item. As described above, the aggregate data of the attributes of the categorical variable is a set of vectors for each item. The decomposition means 4 may decompose this set into individual vectors. The disassembling means 4 can also inherit the label information indicating that the data is attribute data of the categorical variable in the aggregated data (individual vectors for each item) after the disassembly.

  The dimension of the vector obtained as the aggregated data of the attributes of the continuous variable and the dimension of the vector obtained as a result of decomposing the aggregated data of the attributes of the categorical variable for each item are common.

  The clustering means 5 normalizes each vector obtained as aggregate data of continuous variable attributes and a vector obtained as a result of decomposing the aggregate data of categorical variable attributes for each item. The normalization method is not particularly limited. For example, the clustering unit 5 may convert the elements in the vector so that the elements in the vector are in the range of 0 to 1 for each vector. For example, for each vector, the clustering unit 5 may subtract the average value of the elements from the individual elements in the vector and divide each subtraction result by the variance of the elements.

  Then, the clustering unit 5 targets the vector obtained by normalizing the aggregated data of the attributes of the continuous variable and the vector obtained by decomposing the aggregated data of the attributes of the categorical variable for each item and further normalizing. Perform clustering. A cluster obtained as a result of clustering may include a vector to which label information of attribute of continuous variable is added and a vector to which label information of attribute of categorical variable is added.

  The clustering means 5 may cluster each vector by, for example, the K-means method. Note that the clustering method is not limited to the K-means method, and the clustering means 5 may cluster each vector by another method.

  The display means 6 displays the aggregated data (vector subjected to decomposition and normalization) as a heat map for each cluster obtained as a result of clustering. At this time, the display means 6 displays the heat map with the attribute designated to the designation receiving means 2 as the X axis and other attributes and items (category variable attribute items) as the Y axis. The display means 6 displays the attribute name or item name together with the heat map. When label information is added to the data, the display means 6 also displays label information.

  The designation receiving unit 2, the total data generating unit 3, the decomposing unit 4, the clustering unit 5, and the display unit 6 are realized by, for example, a CPU of a computer that performs an input device and a display device. In this case, for example, the CPU reads a display program from a program recording medium such as a program storage device (not shown in FIG. 1) of the computer, and in accordance with the display program, the designation receiving means 2, the total data generating means 3, and the decomposing means 4 The clustering means 5 and the display means 6 may be operated. In addition, regarding the display means 6, the part which makes a display apparatus perform a display is implement | achieved by CPU. Each means may be realized by separate hardware.

  FIG. 3 is a flowchart illustrating an example of processing progress of the display system 1 according to the first embodiment. Assume that a set (for example, see FIG. 2) of a set of attribute values of a plurality of attributes associated with each other is input in advance to the display system as input data.

  First, the designation accepting unit 2 accepts designation of one attribute among a plurality of attributes included in the input data from the operator (step S1). FIG. 4 is a schematic diagram illustrating an example of a display screen of the display system 1 according to the first embodiment. The display screen has a heat map display area 21 and an attribute designation area 22. FIG. 4 shows a state where the heat map is not yet displayed. The attribute designation area 22 is an area for designating an attribute corresponding to the X axis of the heat map. For example, in the attribute designation area 22, each attribute included in the input data is displayed together with a radio button. In this case, when the operator clicks a radio button of an attribute designated, the designation receiving unit 2 determines that the attribute is designated, and decides that attribute as an attribute corresponding to the X axis of the heat map. . For example, the display unit 6 displays the display screen illustrated in FIG. 4 on the display device.

  Next, the total data generation unit 3 creates total data for each attribute other than the specified attribute (the attribute specified in step S1) (step S2).

  As already described, when generating the aggregated data regarding the attribute of the continuous variable, the aggregated data generating unit 3 calculates the statistical value of the attribute value corresponding to the category as the aggregated value for each category of the attribute value of the specified attribute. By doing so, a vector to be aggregated data is generated.

  Further, when generating the aggregated data regarding the attribute of the categorical variable, the aggregated data generating unit 3 stores the category type variable item (attribute value of the categorical variable) corresponding to the category for each attribute value category of the specified attribute. The number is counted, and a vector in which the count values are arranged in the order of division is obtained for each item. The total data generation means 3 determines this set of vectors as total data of the attribute of the categorical variable.

  FIG. 5 is a schematic diagram showing total data of other attributes (each attribute such as “precipitation” illustrated in FIG. 2) when “date” is designated in step S1. The aggregate data is obtained as a vector or a set of vectors. In FIG. 5, the aggregate data is schematically shown in a heat map display mode. Each stage shown in FIG. 5 corresponds to one vector. These points are the same in FIGS. 6 and 7 described later. Since “precipitation”, “temperature”, “number of customers”, and “wind speed” are attributes of continuous variables, the aggregated data is represented by one vector. Since “sales promotion campaign” is an attribute of a categorical variable, the aggregated data is represented by a set of a plurality of vectors (in the example shown in FIG. 5, it is represented by a plurality of stages). FIG. 5 shows an example in which “date” is designated, but the designated attribute (attribute corresponding to the X axis) is not limited to “date”, and may be an arbitrary attribute.

  Next, the decomposing means 4 classifies the aggregate data generated for each attribute into continuous variable aggregate data and categorical variable aggregate data, and decomposes the categorical variable attribute aggregate data for each item. (Step S3). At this time, the decomposing means 4 may add label information indicating whether the data is the aggregated data of the attribute of the continuous variable or the aggregated data of the attribute of the categorical variable to the classified aggregated data. The disassembling means 4 can also inherit the label information indicating the attribute data of the categorical variable in the aggregated data (individual vectors for each item) after the disassembly. Hereinafter, data obtained as a result of disassembling the aggregate data of the attribute of the categorical variable for each item may be referred to as item aggregate data.

  In the example shown in FIG. 5, the aggregate data of the attribute of the categorical variable is a set of vectors obtained for each of seven items such as “festival”, “TV commercial”, “none”,. Therefore, the decomposition unit 4 may decompose the set of vectors into seven vectors corresponding to individual items.

  FIG. 6 is a schematic diagram showing the state after the processing of step S3 is performed on the aggregate data schematically shown in FIG. “P” shown in FIG. 6 is label information indicating that it is aggregated data of attributes of continuous variables. “Q” shown in FIG. 6 is label information indicating attribute data of a categorical variable.

  Next, the clustering means 5 performs clustering on the aggregated data of the attributes of the continuous variables and the aggregated data of each item (Step S4). Specifically, the clustering means 5 first normalizes each vector obtained as the aggregated data of the attributes of the continuous variable and each vector obtained as the aggregated data of each item. Then, the clustering means 5 performs clustering for each vector obtained by normalizing the aggregate data of the attributes of the continuous variable and each vector obtained by normalizing the aggregate data of the individual items. The clustering means 5 may cluster each vector by, for example, the K-means method.

  FIG. 7 is a schematic diagram illustrating a result of performing the clustering process in step S4 on the aggregate data schematically illustrated in FIG. FIG. 7 illustrates a case where the aggregated data of the attributes of the continuous variable and the aggregated data of each item are classified into 7 clusters.

  Next, the display unit 6 displays the aggregated data as a heat map for each cluster (step S5). In step S5, the display means 6 displays the heat map with the attribute designated in step S1 as the X axis (horizontal axis) and other attributes and items as the Y axis (vertical axis). The display means 6 displays the heat map in the heat map display area 21 in the screen (see FIG. 4) displayed on the display device. FIG. 8 is a schematic diagram illustrating an example of a heat map displayed in step S5.

  The display system 1 of the present embodiment clusters the aggregated data of continuous variable attributes and the aggregated data of each item of categorical variable attributes, and displays the aggregated data as a heat map for each cluster obtained as a result. Therefore, an operator of the display system 1 (that is, an observer of the heat map) can easily grasp an attribute having a similar change pattern among a plurality of attributes. Further, the total data of the attribute of the categorical variable is decomposed into total data for each item (attribute value) of the attribute. Therefore, for each item of the attribute of the continuous variable and the attribute of the categorical variable, the operator can easily grasp the attributes and items having similar change patterns.

  FIG. 9 is a diagram illustrating an example of a heat map of the total data displayed for each cluster. When the heat map shown in FIG. 9 is displayed, the operator can easily grasp that the total values of the five attributes shown in FIG. Further, from the heat map of the upper cluster shown in FIG. 9, the operator has only the product A whose sales are proportional to the number of customers, and the products B, C, D... With respect to (not shown), it can be determined that sales are not proportional to the number of customers. Further, for example, the operator can grasp from the heat map of the lower cluster shown in FIG. 9 that the highest discomfort index tends to increase when the humidity and temperature are high.

Embodiment 2. FIG.
FIG. 10 is a block diagram illustrating a configuration example of a display system according to the second embodiment of this invention. The same elements as those in the first embodiment are denoted by the same reference numerals as those in FIG. The display system 1 of the present embodiment includes a selection receiving unit 7 and a narrowing-down unit 8 in addition to the elements in the first embodiment (see FIG. 1).

  The selection accepting unit 7 accepts selection of aggregate data from an operator via an input device (not shown in FIG. 10). Aggregated data that is a selection candidate is aggregated data of attributes of continuous variables other than specified attributes, and aggregated data of items of attributes of categorical variables other than specified attributes.

  FIG. 11 is a schematic diagram illustrating an example of a display screen of the display system 1 according to the second embodiment. The display screen in the second embodiment includes a total data selection area 23 in addition to the heat map display area 21 and the attribute designation area 22. For example, the display unit 6 displays the display screen illustrated in FIG. 11 on the display device. For example, the display unit 6 displays each attribute of the continuous variable other than the designated attribute and each item of the attribute of the categorical variable other than the designated attribute together with the radio button. In this case, when the operator clicks a radio button corresponding to one of the attributes of the continuous variable or one of the items of the attribute of the categorical variable, the total data selection area 23 is Then, it is determined that the total data of the attribute or item corresponding to the clicked radio button is selected.

  The narrowing down means 8 narrows down the aggregate data to be displayed based on the aggregate data selected by the operator. As described in the first embodiment, the aggregate data of the attributes of the continuous variable and the aggregate data of each item of the attribute of the categorical variable are determined as vectors. The narrowing-down means 8 calculates the distance between the vector corresponding to the selected aggregate data and the vector corresponding to the other aggregate data for each aggregate data other than the selected aggregate data, and displays based on the distance. Narrow down the target aggregated data. Here, the vector used when calculating the distance between the vectors is a vector that has been normalized by the clustering means 5.

  The narrowing-down means 8 determines the aggregate data selected by the operator as a display target. Further, the narrowing-down means 8 determines the other aggregate data as a display target on the condition that the distance between the vector corresponding to the selected aggregate data and the vector corresponding to the other aggregate data is equal to or less than a threshold value. . That is, when the distance between the vector corresponding to the selected total data and the vector corresponding to the other total data exceeds the threshold, the other total is excluded from the display target. As a result, the aggregate data to be displayed is narrowed down.

  The display means 6 displays only the aggregated data narrowed down by the narrowing-down means 8 (that is, only the aggregated data determined as a display target) as a heat map.

  FIG. 12 is a schematic diagram illustrating a display example of the heat map when the aggregate data to be displayed is narrowed down after the heat map illustrated in FIG. 8 is displayed. By narrowing down the aggregate data to be displayed, the number of attributes and items displayed as a heat map is reduced. In the example shown in FIG. 12, the total data belonging to the clusters 1, 2, and 5 shown in FIG. 8 are all displayed. However, as a result of narrowing down the aggregated data, some aggregated data belonging to the cluster before narrowing down may be excluded from the display target.

  Further, the above-described threshold value may be a predetermined fixed value. Or the structure which an operator can select a threshold value may be sufficient. In the example illustrated in FIG. 11, the slider 24 is displayed in the total data selection area 23 as a GUI (Graphic User Interface) for the operator to select a threshold value. The selection accepting means 7 may accept the selection of the threshold value via the slider 24 together with the selection of the total data. Note that the GUI for selecting the threshold value may not be the slider 24. When the threshold is a fixed value, the GUI for selecting the threshold need not be displayed.

  The selection receiving means 7 and the narrowing-down means 8 are realized by a CPU that operates according to a display program, for example. Each means including the selection receiving means 7 and the narrowing-down means 8 may be realized by separate hardware.

  The display system 1 of the second embodiment narrows down the aggregate data similar to the aggregate data selected by the operator and displays it as a heat map. Accordingly, the operator can more easily grasp attributes and items whose change pattern is similar to the aggregate data similar to the aggregate data of interest.

  FIG. 13 is a schematic block diagram illustrating a configuration example of a computer according to each embodiment of the present invention. The computer 1000 includes a CPU 1001, a main storage device 1002, an auxiliary storage device 1003, an interface 1004, a display device 1005, and an input device 1006. The input device 1006 is, for example, a radio button or a mouse for operating the slider 24.

  The display system 1 described above is mounted on the computer 1000. The operation of the display system 1 is stored in the auxiliary storage device 1003 in the form of a program (display program). The CPU 1001 reads out the program from the auxiliary storage device 1003, develops it in the main storage device 1002, and executes the above processing according to the program.

  The auxiliary storage device 1003 is an example of a tangible medium that is not temporary. Other examples of the tangible medium that is not temporary include a magnetic disk, a magneto-optical disk, a CD-ROM, a DVD-ROM, and a semiconductor memory connected via the interface 1004. When this program is distributed to the computer 1000 via a communication line, the computer 1000 that has received the distribution may develop the program in the main storage device 1002 and execute the above processing.

  The program may be for realizing a part of the above-described processing. Furthermore, the program may be a differential program that realizes the above-described processing in combination with another program already stored in the auxiliary storage device 1003.

  Next, the minimum configuration of the present invention will be described. FIG. 14 is a block diagram showing an example of the minimum configuration of the display system of the present invention. The display system of the present invention includes total data generation means 53, clustering means 55, and display means 56.

  The aggregate data generation means 53 (for example, the aggregate data generation means 3) is given a set of attribute value pairs of a plurality of attributes associated with each other, and when one attribute is designated among the plurality of attributes, Aggregated data obtained by aggregating attribute values of attributes that are not specified according to attribute value classifications of the specified attributes is generated for each attribute that is not specified.

  Clustering means 55 (for example, clustering means 5) clusters the aggregated data.

  The display unit 56 (for example, the display unit 6) displays the aggregated data as a heat map for each cluster obtained as a result of clustering using the designated attribute as the X axis.

  With such a configuration, an operator of the display system (in other words, an observer of the heat map) can easily grasp an attribute having a similar change pattern among a plurality of attributes.

  In addition, the aggregate data generation unit 53 calculates the statistical value of the corresponding attribute value as the aggregate value for each attribute value category for the attribute of the continuous variable among the attributes that are not specified. Among the attributes that are not specified, for categorical variable attributes, the count value of each item that is the attribute value of the corresponding categorical variable attribute is aggregated for each attribute value category of the specified attribute. A decomposing unit (e.g., decomposing unit 4) that decomposes the categorical variable attribute for each item, and the clustering unit 55 calculates the categorical variable attribute total data after decomposing, A configuration may be used in which the aggregated data of the attributes of the continuous variable is clustered.

  Moreover, the structure provided with the designation | designated reception means (for example, designation | designated reception means 2) which receives designation | designated of one attribute among several attributes may be sufficient.

  In addition, the configuration includes a selection receiving unit (for example, a selection receiving unit 7) that receives selection of total data, and a narrowing unit (narrowing unit 8) that narrows down the total data to be displayed based on the selected total data. There may be.

  Further, the narrowing-down means may be configured to narrow down the aggregate data to be displayed based on the distance between the vector represented as the selected aggregate data and each vector represented as other aggregate data.

  In addition, the narrowing down means that the selected aggregated data is set as a display target, and the distance between the vector represented as the selected aggregated data and the vector represented as other aggregated data is equal to or less than a threshold. In addition, the other aggregated data may be determined as a display target.

  Further, the selection accepting unit may be configured to accept selection of a threshold value together with selection of aggregate data.

  The present invention is suitably applied to a display system that displays a heat map.

DESCRIPTION OF SYMBOLS 1 Display system 2 Specification reception means 3 Total data generation means 4 Decomposition means 5 Clustering means 6 Display means 7 Selection reception means 8 Refinement means

Claims (7)

When a set of attribute value pairs of a plurality of attributes that are associated with each other is given, and one attribute of the plurality of attributes is specified, the attribute values of the attributes that are not specified are Aggregated data generating means for generating aggregated data aggregated according to the attribute value category for each of the attributes not specified;
Clustering means for clustering the aggregated data;
Display means for displaying the aggregated data as a heat map for each cluster obtained as a result of clustering using the designated attribute as the X axis ;
Selection accepting means for accepting selection of aggregate data;
And a narrowing means for narrowing down the aggregate data to be displayed based on the selected aggregate data,
The narrowing means is
Narrow down the aggregate data to be displayed based on the distance between the vector represented as the selected aggregate data and each vector represented as other aggregate data
Display system comprising a call. The aggregate data generation means
Among the attributes that are not specified, for continuous variable attributes, the statistical value of the corresponding attribute value is calculated as an aggregate value for each attribute value category of the specified attribute,
Among the attributes that are not specified, for categorical variable attributes, the count value of each item that is the attribute value of the corresponding categorical variable attribute is aggregated for each category of the attribute value of the specified attribute. As a value,
Disassembling means for decomposing the aggregate data of the attribute of the categorical variable for each item,
Clustering means
The display system according to claim 1, wherein the aggregated data of the attribute of the categorical variable after decomposition and the aggregated data of the attribute of the continuous variable are clustered. The display system according to claim 1, further comprising designation receiving means for accepting designation of one of a plurality of attributes. Narrowing means
The selected aggregate data is defined as a display target, and the other data is provided on the condition that the distance between the vector represented as the selected aggregate data and the vector represented as the other aggregate data is equal to or less than a threshold value. Set the aggregate data as the display target
The display system according to any one of claims 1 to 3 . The display system according to claim 4 , wherein the selection accepting unit accepts selection of a threshold value as well as selection of aggregate data. When a set of attribute value pairs of a plurality of attributes that are associated with each other is given, and one attribute of the plurality of attributes is specified, the attribute values of the attributes that are not specified are Generate aggregate data aggregated according to the attribute value category for each of the attributes not specified,
Clustering the aggregated data;
The aggregated data is displayed as a heat map for each cluster obtained as a result of clustering using the designated attribute as the X axis ,
Accept selection of aggregate data,
Based on the selected aggregate data, filter the aggregate data to be displayed,
When narrowing down the aggregate data to be displayed, narrow down the aggregate data to be displayed based on the distance between the vector represented as the selected aggregate data and each vector represented as other aggregate data. Characteristic display method. On the computer,
When a set of attribute value pairs of a plurality of attributes that are associated with each other is given, and one attribute of the plurality of attributes is specified, the attribute values of the attributes that are not specified are Aggregated data generation processing for generating aggregated data aggregated according to attribute value categories for each of the attributes not specified,
A clustering process for clustering the aggregated data ;
Display processing for displaying the aggregated data as a heat map for each cluster obtained as a result of clustering using the designated attribute as the X-axis ,
A process for accepting selection of aggregate data, and
Based on the selected aggregate data, the process to narrow down the aggregate data to be displayed is executed,
In the process of narrowing down the aggregated data to be displayed, the aggregated data to be displayed is narrowed down based on the distance between the vector represented as the selected aggregated data and each vector represented as other aggregated data
Because of the display program.

Images