Analysis System

Abstract

Properly analyze campaigns conducted through social media. [Solution] An analysis system 1 for analyzing campaign-related information related to a campaign using social media, comprising a processor 3 and a memory 4 used by the processor, wherein the following functions are realized when the processor executes a predetermined computer program stored in the memory; a collection analysis code generation unit 11 that generates a collection analysis code for analyzing campaign-related information collected from social media based on an input document 6; and a collection analysis execution unit 14 that collects campaign-related information from social media based on the generated collection analysis code and calculates a score for evaluating the campaign based on the collected campaign-related information.

Description

Technical Field

[0006] , , , , , , , , , , ,

[0001] The present invention relates to an analysis system.

Background Art

[0002] There are various applications for analyzing campaigns using social media. For example, in the analysis of a brand's campaign, it is analyzed how viral the brand is and whether the posts are from non-fake human accounts. Furthermore, due to the emergence of new services, specification changes, and different usage methods depending on the region, social media requires various analyses.

[0003] Patent Document 1 describes a system for quantitatively analyzing the virality (mediacy) of a brand from the content of social media. This system analyzes the attributes of the poster, the quality of the poster (such as not being fake), the mediacy of the posted content, engagement (such as the number of likes), and the reliability of the post (such as the relationship with the brand), and calculates a score.

[0004] Non-Patent Document 1 describes providing a service for visualizing the effect of influencers on social media and enhancing awareness and engagement with customers.

Prior Art Documents

Patent Documents

[0005]

Patent Document 1

Non-Patent Documents

[0006]

Non-Patent Document 1

Summary of the Invention

[0007] Analyzing campaigns conducted through social media requires flexibility because social media itself is constantly changing. Furthermore, because social media posts and their dissemination are faster than traditional media such as newspapers and television, speed of analysis is also essential.

[0008] Therefore, the present invention has been made in view of the above problems, and its purpose is to provide a technology for appropriately analyzing campaigns conducted through social media. [Means for solving the problem]

[0009] Therefore, the present invention has been made in view of the above problems, and is an analysis system for analyzing campaign-related information related to a campaign using social media, comprising a processor and a memory used by the processor, wherein the following functions are realized when the processor executes a predetermined computer program stored in the memory; a collection analysis code generation unit that generates a collection analysis code for analyzing the campaign-related information collected from the social media based on an input document; and a collection analysis execution unit that collects the campaign-related information from the social media based on the generated collection analysis code and calculates a score for evaluating the campaign based on the collected campaign information. [Effects of the Invention]

[0010] According to the present invention, campaigns conducted through social media can be appropriately analyzed. [Brief explanation of the drawing]

[0011] [Figure 1] A block diagram showing the overall configuration of the analysis system. [Figure 2] A diagram showing an example of a data collection and analysis code. [Figure 3] A diagram showing an example of a countermeasure planning code. [Figure 4] A diagram showing an example of a countermeasure execution transition table. [Figure 5] A diagram showing an example of a constraint condition parameter. [Figure 6] A flowchart showing an example of the analysis process up to the countermeasure simulation. [Figure 7] A diagram showing an example of the collected analysis code generation process. [Figure 8] A diagram showing an example of the simulation process. [Figure 9] An interface diagram showing an example of a countermeasure simulation screen. [Figure 10] A flowchart showing an example of the analysis process after information collection. [Figure 11] A diagram showing an example of social media data. [Figure 12] An interface diagram showing an example of a monitoring dashboard screen.

Mode for Carrying Out the Invention

Examples

[0012] Hereinafter, a specific example of the analysis system according to the embodiment of the present invention will be described while referring to the drawings. Note that the present invention is not limited by the examples, but is shown by the scope of the claims. ​​​​​​​​​ The computer 2 includes a CPU 3, a memory 4, and a communication I / F 5 as an example of a "processor". The document 6 and the constraint parameter 7 are input into the computer 2 by the user of the analysis system 1. The document 6 is a human-readable document written in natural language. The document 6 includes the analysis of past campaigns, the content of specification changes related to the social media 8, and the content of differences in usage methods of the social media 8 for each region. The social media 8 is the target for the computer 2 to collect social media data through crawling via the communication I / F 5 or API access. The action target 9 of the Whole-of-Society approach is the entity that executes countermeasure actions according to instructions from the computer 2, specifically, the government, non-governmental organizations, citizens, etc.

[0016] In the memory 4, the processes of the large language model unit 10, the collection analysis code generation unit 11, the countermeasure formulation code generation unit 12, the countermeasure simulation unit 13, the collection analysis execution unit 14, the countermeasure execution instruction unit 15, and the countermeasure evaluation unit 16 are executed.

[0017] The large language model unit 10 reads the input document 6 and analyzes the content related to the collection analysis code and the content related to the countermeasure formulation code. The collection analysis code generation unit 11 newly creates a collection analysis code from the content of the document 6 or identifies and corrects the modified parts of the existing collection analysis code.

[0018] The countermeasure formulation code generation unit 12 newly creates a countermeasure formulation code from the content of the document 6 or identifies and corrects the modified parts of the existing countermeasure formulation code. The countermeasure simulation unit 13 provides an interface for inputting the constraint parameter 7, comprehensively calculates the combinations of countermeasures derived from the countermeasure formulation code, and evaluates the ranking of satisfying the constraints.

[0019] The data collection and analysis execution unit 14 collects and analyzes social media data from social media 8 based on the data collection and analysis code and calculates a score. Specifically, the data collection and analysis execution unit 14 collects campaign information related to a campaign from social media 8 and calculates a score to evaluate the campaign based on the collected campaign information. The countermeasure execution instruction unit 15 instructs the Whole-of-Society approach action target 9 countermeasures based on the score result, based on the countermeasure planning code. The countermeasure evaluation unit 16 evaluates the appropriateness of the countermeasure instruction and displays the evaluation result to the user of computer 2.

[0020] Here, computer 2 may be a standalone server or a virtual machine running on a cloud service.

[0021] Figure 2 shows an example of a data collection and analysis code.

[0022] The data collection and analysis code 100 has a tree-like structure in which input nodes 101, 102, preprocessing nodes 103, 104, analysis nodes 105-107, and score nodes 108-110 are connected by a pipeline. Input codes 101 and 102 contain program code for collecting data from social media 8 by crawling or API access. Preprocessing nodes 103 and 104 contain program code for normalizing the data or converting it to an appropriate data type. Analysis nodes 105-107 contain program code for analyzing the data based on predetermined perspectives. These predetermined perspectives include attributes of the poster (engagement rate, ratio of real to fake followers), visibility of the post content (number of views, number of repeats), engagement attributes (number of likes, number of shares), and trust attributes (hashtags used, image content). Score nodes 108-110 contain program code for calculating a score that takes a value from 0 to 100 based on the results of analysis nodes 105-107. By connecting these nodes in a pipeline, it becomes easy to specify or rearrange the sequence of processes that follow a given process.

[0023] Figure 3 shows an example of a countermeasure planning code.

[0024] The countermeasure planning code 200 has a graph structure in which transitions 201-204, which are triggers for state transitions, and places 210-215, which indicate states, are connected by arcs. Each of the transitions 201-204 has an identifier 205 and a content 206. For example, transition T1(201) is a state transition from place S(210) to place P1(211) when the condition "score 1 > α" is met. Each of the transitions 210-215 has an identifier 216, a content 217, quality 218, cost 219, and delivery date 220. For example, place P1(211) is one of the action targets 9 of the Whole-of-Society approach, which is to implement countermeasure 1, and the quality 218, cost 219, and delivery date 220 associated with that implementation are represented numerically. A high quality 218 indicates that the countermeasures have a direct effect, while a low quality 218 indicates that the countermeasures have an indirect effect. A high cost 219 indicates that the cost of the countermeasures is high, while a low cost 219 indicates that the cost of the countermeasures is low. A short delivery time 220 means that there is little time to prepare for the implementation of the countermeasures, while a long delivery time 220 means that there is a long preparation period.

[0025] In addition to representing state transitions, the countermeasure planning code 200 can also represent branching, parallel processing, hierarchical structures, and time constraints. These state transitions, branching, parallel processing, hierarchical structures, and time constraints are known as Petri nets and their applications. Furthermore, they can also be obtained by converting from other models, such as Business Process Model and Notation (BPMN).

[0026] Figure 4 shows an example of a countermeasure implementation transition table.

[0027] The countermeasure execution transition table 300 has a table structure consisting of multiple rows, with each row containing a countermeasure execution transition 301, quality 302, cost 303, delivery date 304, and ranking 305 as item values ​​(column values). Countermeasure execution transition 301 stores all possible state transitions when a token is fired on the Petri net of the countermeasure planning code 200. Quality 302, cost 303, and delivery date 304 store the quality, cost, and delivery date, respectively, when countermeasure execution transition 301 is taken, as numerical values. Ranking 305 is the number obtained when the quality of countermeasure execution transition 301 is sorted in descending order.

[0028] Figure 5 shows an example of constraint parameters.

[0029] The constraint parameter 7 stores the ideal values ​​314 and acceptable values ​​315 for quality 311, cost 312, and delivery time 313, respectively. The ideal value 314 is the ideal value for the countermeasure execution transition 301. The acceptable value 315 is the acceptable value for the countermeasure execution transition 301.

[0030] Figure 6 is a flowchart showing an example of the analysis process leading up to the countermeasure simulation.

[0031] Computer 2 performs analysis processing of campaigns conducted via social media 8. First, the large-scale language model unit 10 reads document 6 (S401). Document 6 is an image file containing an office file, a web page, or a book.

[0032] Next, the large-scale language model unit 10 analyzes document 6 (S402). If document 6 is a report analyzing a campaign by social media 8, the large-scale language model unit 10 identifies the parts related to data collection and analysis, and the planning of countermeasures in the analysis process. For example, if there are differences in the analysis process depending on the region, the large-scale language model unit 10 identifies those differences. If document 6 is a service specification for a new social media 8, the large-scale language model unit 10 identifies the parts related to API access methods and the format of available data. If document 6 is a notification regarding specification changes for an existing social media 8, the large-scale language model unit 10 identifies the parts that have been changed due to the specification changes. The analysis by the large-scale language model unit 10 is not limited to the above, but also includes analysis that contributes to the creation and modification of data collection and analysis code 100 or countermeasure planning code 200.

[0033] Next, the data collection and analysis code generation unit 11 creates a new data collection and analysis code 100 or modifies an existing data collection and analysis code 100 based on the information obtained by the large-scale language model unit 10 (S403). An example of the operation of the data collection and analysis code 11 (data collection and analysis code generation process) is illustrated using Figure 7.

[0034] Figure 7 shows an example of the data collection and analysis code generation process.

[0035] Figure 7(a) shows an example in which the simplest collection and analysis code 100 is newly created by the collection and analysis code generation unit 11. The collection and analysis code generation unit 11 creates the collection and analysis code 100 which includes an input node 511, a preprocessing node 512, an analysis node 513, and a score node 514.

[0036] Figures 7(b) and 7(c) show an example of the data collection and analysis code generation unit 11 modifying an existing data collection and analysis code 100. In Figure 7(b), the data collection and analysis code generation unit 11 identifies a location 517 where a new analysis will be performed following the preprocessing node. In Figure 7(c), the data collection and analysis code generation unit 11 adds a new analysis node 515 and a new score node 516 to the location 517 where the new analysis will be performed.

[0037] The countermeasure planning code generation unit 12 creates a new countermeasure planning code 200 or modifies an existing countermeasure planning code 200 based on the information obtained by the large-scale language model unit 10 (S404).

[0038] The countermeasure simulation unit 13 performs a simulation of the implementation of the countermeasure based on the countermeasure planning code 200. Figure 8 shows an example of the operation (simulation process) of the countermeasure simulation unit 13 (S405).

[0039] Figure 8 shows an example of the simulation process.

[0040] Figure 8(a) shows the initial state based on countermeasure planning code 200. In the initial state, token 528 exists at place S(523).

[0041] Figure 8(b) shows that when token 528 is in place S(523), transition T1(521) is triggered, and token 528 transitions to place P1(524).

[0042] Figure 8(c) shows that when token 528 is in place P1(524), transition T2(522) is triggered, causing token 528 to transition to place P2(525) and then to place P3(526).

[0043] It is known that all state transitions in a Petri net can be identified through such simulations. The countermeasure simulation unit 13 stores all countermeasure execution transitions in the countermeasure execution transition table 300. Furthermore, the countermeasure simulation unit 13 calculates quality 302, cost 303, and delivery time 304. Under the constraint parameter 7, the countermeasure simulation unit 13 calculates a recommended ranking of countermeasure execution transitions 301 that are suitable for satisfying the constraints, and a ranking of non-recommended transitions that do not satisfy the constraints but would be unrecommended if they were possible.

[0044] The countermeasure simulation unit 13 displays the simulation results (S406).

[0045] Figure 9 is an interface diagram showing an example of a countermeasure simulation screen.

[0046] The countermeasure simulation screen 500 includes a data collection and analysis code editing field 501, a countermeasure planning code editing field 502, and a countermeasure simulation results field 503.

[0047] In the data collection and analysis code editing area 501, the input node 511, preprocessing node 512, analysis nodes 513 and 515, and score nodes 514 and 516 are displayed as clickable icons. Clicking on any node makes the program code editable. Pipelines between the nodes can be connected, disconnected, and reconnected using drag-and-drop.

[0048] In the countermeasure planning code editing field 502, transitions 521 and 522, and places 523, 524, 525, and 526 are displayed as clickable icons. When transitions 521 and 522 are clicked, their attributes, namely identifier 205 and content 206, become editable. When places 523, 524, 525, and 526 are clicked, their attributes, namely identifier 216, content 217, quality 218, cost 219, and delivery date 220, become editable.

[0049] The countermeasure simulation results column 503 displays a table consisting of multiple rows, with each row representing a countermeasure execution transition 531, quality 532, cost 533, delivery time 534, and ranking 535. In the countermeasure simulation results column 503, among the rankings 535, countermeasure execution transitions 531 that satisfy constraint parameter 7 are highlighted with a thick border 536.

[0050] Figure 10 is a flowchart showing an example of the analysis process after information collection.

[0051] The data collection and analysis execution unit 14 collects social media data (see Figure 11) by crawling or accessing social media 8 via API (S601).

[0052] Figure 11 shows an example of social media data.

[0053] The social media data 700 has a table structure consisting of multiple entries, each containing SNS type 701, time 702, account identifier 703, post content 704, hashtag 705, and URL 706 as a single entry. The post content 704 is not limited to text; it may also be audio, images, or videos. The hashtag 705 is the label of the post included in the post content 704. The URL 706 is the URL included in the post content 704. The social media data 700 becomes the data read by the input nodes 101 and 102 of the collection and analysis code 100.

[0054] The data collection and analysis execution unit 14 performs analysis processing based on the data collection and analysis code 100 and calculates a score (S602).

[0055] The countermeasure execution instruction unit 15 issues an instruction to execute the countermeasure based on the countermeasure planning code 200 and the score calculated in S602 (S603). When the countermeasure is executed, it will affect social media 8, and the analysis system 1 issues an instruction to execute a new countermeasure by calculating a new score.

[0056] The countermeasure evaluation unit 16 calculates a ranking of the most recent countermeasure execution instructions (S604). The ranking is based on the best possible move within the range allowed by the constraint parameter, and is also better if it is not the worst possible move within the range that is not allowed.

[0057] The countermeasure evaluation unit 16 displays the results of the countermeasure implementation on the monitoring dashboard (S605).

[0058] Figure 12 is an interface diagram showing an example of a monitoring dashboard screen.

[0059] The monitoring dashboard screen 800 includes a collection section 801, an analysis section 802, and a countermeasures section 803.

[0060] The collection column 801 displays the collection status of social media data 700, including SNS type 811, number of items 812, and a time-series graph 813.

[0061] Analysis section 802 displays scores 821, 822, and 823, which are the results of analyzing social media data 700. For example, scores 821 through 823 can represent one of the following: virality, engagement, or poster quality (e.g., whether the content is fake).

[0062] The countermeasures section 803 displays the most recent countermeasures instruction 831 and the ranking 832. The countermeasures instruction 831 can be, for example, "Instruct countermeasures again," "Instruct different countermeasures," or "Leave it to word of mouth."

[0063] According to this configuration, the analysis system 1 analyzes campaign-related information related to a campaign using social media 8 as a medium, and comprises a processor 3 and a memory 4 used by the processor 3, wherein the following functions are realized when the processor 3 executes a predetermined computer program stored in the memory 4: a collection analysis code generation unit 11 that generates a collection analysis code 100 for analyzing campaign-related information collected from social media 8 based on an input document 6, and a collection analysis execution unit 14 that collects campaign-related information from social media 8 based on the generated collection analysis code 100 and calculates a score for evaluating the campaign based on the collected campaign-related information.

[0064] This enhances the speed and flexibility of campaign analysis using social media as a source of information, enabling proper analysis of campaigns conducted through social media. Furthermore, it reduces the cost of analysis. Additionally, a series of countermeasures can be implemented systematically and reproducibly. Moreover, it can be used for campaign simulations for training purposes.

[0065] Document 6 includes at least one of the following: changes to the Social Media 8 specifications, regional usage, or past campaign-related information. This allows for a more appropriate analysis of campaigns using Social Media 8, even if the functionality of Social Media 8 changes, and especially if the campaign is repeated, in accordance with regional usage of Social Media 8.

[0066] The system includes a countermeasure planning code generation unit 12 that generates a countermeasure planning code 200 that instructs the execution of countermeasures according to the score based on the input document 6, and a countermeasure simulation unit 13 that simulates the transition of countermeasure execution that satisfies predetermined constraints on the countermeasure planning code 200. This makes it possible to instruct the execution of countermeasures according to the score within the range that satisfies the constraints.

[0067] It should be noted that the present invention is not limited to the embodiments described above, and various modifications are included. For example, the embodiments described above are described in detail to make the present invention easier to understand, and are not necessarily limited to those having all the configurations described. Furthermore, it is possible to replace parts of the configuration of one embodiment with the configuration of another embodiment, and it is also possible to add configurations from other embodiments to the configuration of one embodiment. In addition, it is possible to add, delete, or replace parts of the configuration of each embodiment with other configurations. [Examples]

[0068] The analysis system 1 in Example 2 will be described with reference to Figure 1 again. Example 2 shows an alternative method for creating the collection analysis code 100 and the action planning code 200. For example, consider analyzing a social media 8 campaign for a regularly occurring event, such as an annual event. Using the document 8 from the previous event as input, the collection analysis code 100 and the action planning code 200 for the current event are generated. By creating these codes, it is possible to compare whether the scores observed during the previous event are also observed during the current event.

[0069] According to this, the analysis of social media campaigns for events that occur every predetermined year can be standardized and streamlined. [Examples]

[0070] The analysis system 1 in Example 3 will be described with reference to Figure 1 again. Example 3 shows yet another method for creating the collection analysis code 100 and the action planning code 200. For example, suppose we are analyzing social media 8 campaigns for similar events in region A and region B. Using document 8 from the event in region A as input, we create the collection analysis code 100 and the action planning code 200 for region B. By creating these codes, we can compare whether the scores observed during the event in region A are also observed during the event in region B.

[0071] This allows for the standardization and streamlining of the analysis of social media campaigns for events occurring in multiple regions. [Explanation of symbols]

[0072] 1…Analysis system, 3…CPU, 4…Memory, 6…Documents, 7…Constraint parameters, 8…Social media, 11…Collection and analysis code generation unit, 12…Countermeasure planning code generation unit, 13…Countermeasure simulation unit, 14…Collection and analysis execution unit, 16…Countermeasure evaluation unit, 100…Collection and analysis code, 200…Countermeasure planning code

Claims

1. An analytical system for analyzing campaign-related information related to campaigns conducted via social media, The system comprises a processor and memory used by the processor, The following functions are realized when the processor executes a predetermined computer program stored in the memory: A collection analysis code generation unit generates a collection analysis code that analyzes the campaign-related information collected from the social media based on the input document. An analysis system comprising: a collection and analysis execution unit that collects campaign-related information from social media based on the generated collection and analysis code, and calculates a score for evaluating the campaign based on the collected campaign-related information.

2. The aforementioned document includes at least one of the following: changes in the specifications of the social media, regional usage, or past campaign-related information. The analysis system according to claim 1.

3. A countermeasure planning code generation unit generates a countermeasure planning code that instructs the execution of countermeasures according to the score based on the input document, The system includes a countermeasure simulation unit that simulates the transition of countermeasure execution that satisfies predetermined constraints on the countermeasure planning code. The analysis system according to claim 1.

4. The aforementioned data collection and analysis code has an input node, a preprocessing node, an analysis node, and a score node connected in a pipeline. The analysis system according to claim 1.

5. The aforementioned countermeasure planning code includes a transition indicating whether the score exceeds a predetermined threshold, and a place that instructs the execution of the countermeasure. The analysis system according to claim 3.

6. The aforementioned place has at least one of the following attributes: quality, cost, or delivery time of the countermeasures implemented according to the score. The analysis system according to claim 5.

7. The countermeasure simulation unit calculates, based on the input constraint parameters, the degree of recommendation when the transition to executing the countermeasure satisfies the constraints, or the degree of deprecation when the transition to executing the countermeasure does not satisfy the constraints. The analysis system according to claim 3.

8. The system includes a countermeasure evaluation unit that evaluates the countermeasures implemented according to the score within the acceptable range of the constraint parameter. The analysis system according to claim 1.

Images