Data analysis device, data analysis method, data analysis program, and data analysis system

Abstract

This data analysis device receives request data that is expressed with natural language and that requests data analysis related to a manufacturing industry. The data analysis device inputs the received request data to a generation model trained in advance to acquire answer data output from the generation model. The data analysis device: executes data analysis indicated by the request data by using data included in the answer data, namely analysis tool data indicating the type of analysis tool to be used when executing the data analysis indicated by the request data, instruction data directed to the analysis tool, and analysis target data indicating the data to be analyzed among a plurality of types of data related to the manufacturing industry. As a result of executing such data analysis, the data analysis device acquires a data analysis result corresponding to the request data, and outputs the data analysis result.

Description

Data analysis device, data analysis method, data analysis program, and data analysis system

[0001] The present disclosure relates to a data analysis device, a data analysis method, a data analysis program, and a data analysis system.

[0002] A computer-operated method for managing a plurality of drawing data representing a set of parts is known (for example, Japanese Patent No. 7413445). This method includes the steps of: determining whether target drawing data included in the plurality of drawing data is related to existing drawing data that has already been stored; acquiring performance information linked to the existing drawing data when the result of the determination is affirmative; and transmitting performance display information for displaying the performance information to a terminal, wherein the determination includes determining whether the existing drawing data represents a part having the same attribute as the part represented by the target drawing data, which attribute is material, surface treatment, processing method, or a combination of two or more of these.

[0003] Also, a technology that uses a large-scale language model to analyze patent information and provides new value is known (for example, Japanese Patent Publication No. 7421740). A patent information analysis program corresponding to this technology causes a computer to execute the following steps: acquiring technical information including a technical idea; acquiring patent information to be compared with the technical information; specifying the technical information and patent information in response to a predetermined instruction statement for evaluating the relevance of the two pieces of information from a specific perspective; and outputting an evaluation result indicating the relevance between the technical information and the patent information in terms of the perspective based on a result obtained from the large-scale language model by inputting the instruction statement specifying the technical information and the patent information into the large-scale language model.

[0004] Also known is an architectural drawing creation support system suitable for adapting editable edit elements in architectural drawings to the specifications of a finishing schedule (see, for example, Japanese Patent No. 7341581). A drawing creation support device 100 in this architectural drawing creation support system inputs a request to a large-scale language model, including finishing schedule data and edit elements, and a request to generate adapted element information for adapted elements that conform to the specifications of the finish elements and are included in the architectural drawing. The drawing creation support device 100 then acquires adapted element information output from the large-scale language model in response to the request, and if the acquired edit elements do not match the adapted elements, displays a message indicating that the edit elements are incomplete.

[0005] However, it is difficult to easily analyze data related to the manufacturing industry. For example, data related to the manufacturing industry may include multiple types of data, such as image data (e.g., drawing data, etc.), document data (e.g., text in drawings or various specifications and reports related to manufacturing, etc.), and numerical data (e.g., sales data of manufactured products or dimensional data in drawings, etc.). In such a situation, for example, when a user wants to analyze such data related to the manufacturing industry to obtain analysis results, it is often difficult to immediately know which data to analyze and how to analyze them to obtain the desired analysis results.

[0006] The technologies disclosed in the above Patent Publications No. 7413445, No. 7421740, and No. 7341581 do not take into account the above-mentioned issues that exist when analyzing data related to the manufacturing industry.

[0007] The present disclosure has been made in consideration of the above circumstances, and aims to easily perform data analysis when data related to the manufacturing industry is composed of multiple types of data.

[0008] In order to achieve the above object, a first aspect of the present disclosure is a data analysis device including: a receiving unit that receives request data requesting data analysis related to the manufacturing industry and that is expressed in natural language; a response data acquisition unit that inputs the received request data into a pre-trained generation model and thereby acquires response data output from the generation model; and an analysis unit that performs the data analysis represented by the request data using analysis tool data that indicates the type of analysis tool to be used when performing the data analysis represented by the request data, which is included in the response data, instruction data for the analysis tool, and analysis target data that indicates the data to be analyzed among multiple types of data related to the manufacturing industry, thereby acquiring data analysis results corresponding to the request data and outputting the data analysis results.

[0009] Furthermore, a second aspect of the present disclosure is a data analysis method in which a computer executes a process of accepting request data that requests data analysis related to the manufacturing industry and that is expressed in natural language, inputting the accepted request data into a pre-trained generative model to obtain response data output from the generative model, and executing the data analysis represented by the request data using analysis tool data that indicates the type of analysis tool to be used when performing the data analysis represented by the request data, instruction data for the analysis tool, and analysis target data that indicates the data to be analyzed from multiple types of data related to the manufacturing industry, thereby obtaining data analysis results corresponding to the request data, and outputting the data analysis results.

[0010] Furthermore, a third aspect of the present disclosure is a data analysis program for causing a computer to execute a process of receiving request data that requests data analysis related to the manufacturing industry and that is expressed in natural language, inputting the received request data into a pre-trained generative model to obtain response data output from the generative model, and executing the data analysis represented by the request data using analysis tool data that indicates the type of analysis tool to be used when performing the data analysis represented by the request data, instruction data for the analysis tool, and analysis target data that indicates the data to be analyzed from multiple types of data related to the manufacturing industry, which are included in the response data, to obtain data analysis results corresponding to the request data, and outputting the data analysis results.

[0011] According to the present disclosure, when data related to the manufacturing industry is configured to include multiple types of data, an effect can be obtained in which data analysis can be easily performed.

[0012] FIG. 1 is a diagram showing an example of a schematic configuration of a data analysis system of this embodiment. FIG. 2 is a diagram showing an example of data related to the manufacturing industry. FIG. 3 is an example of a screen displayed on the display unit of a user terminal. FIG. 4 is an example of a screen displayed on the display unit of a user terminal. FIG. 5 is an example of a screen displayed on the display unit of a user terminal. FIG. 6 is an example of a screen displayed on the display unit of a user terminal. FIG. 7 is a schematic block diagram of a computer functioning as each device in the data analysis system. FIG. 8 is a diagram for explaining processing executed by the data analysis system of this embodiment.

[0013] Hereinafter, the embodiments will be described in detail with reference to the drawings.

[0014] <Data analysis system configuration>

[0015] Fig. 1 is a block diagram showing a data analysis system 10 according to this embodiment. As shown in Fig. 1, the data analysis system 10 according to this embodiment includes a data analysis device 16 and multiple user terminals 18A, 18B, and 18C. In the following, a single user terminal will be referred to as a user terminal 18 unless a specific terminal is being referred to. The data analysis device 16 and the user terminal 18 are communicatively connected via a network 19 such as the Internet.

[0016] The data analysis system 10 of this embodiment easily performs data analysis when data related to the manufacturing industry is configured to include multiple types of data.

[0017] FIG. 2 shows an example of data related to the manufacturing industry. As shown in FIG. 2, for example, if the manufacturing industry is related to metal processing, multiple types of data are analyzed, such as the ID and price of the metal processed product, the drawing number of the drawing depicting the metal processed product, the name, material, surface finish, maximum dimensions, thickness, and processing category. Note that the data shown in FIG. 2 is merely an example, and various types of data are conceivable as multiple types of data related to the manufacturing industry. For example, multiple types of data related to the manufacturing industry may include manufacturing document data, which is document data related to objects displayed in drawing data; manufacturing numerical data, which is numerical data related to objects displayed in drawing data; drawing document data, which is document data present in drawing data; and drawing numerical data, which is numerical data present in drawing data. Thus, when data related to the manufacturing industry is composed of multiple types of data, it is difficult to perform data analysis.

[0018] Therefore, the data analysis system 10 of this embodiment performs data analysis of data related to the manufacturing industry by utilizing a large-scale language model, which is an example of a pre-trained generative model.

[0019] 3 to 7 are diagrams illustrating an example of the operation of the data analysis system 10 of this embodiment. As shown in Fig. 3, the data analysis system 10 displays an interactive screen S on the display unit of the user terminal 18 for communication with the user.

[0020] The user operates his / her user terminal 18 to input a message, which is request data expressed in natural language, into the input field B1 and presses the "Send" button. For example, the user inputs the message "Please create a price distribution" as shown in FIG. 1 into the input field B1.

[0021] The data analysis device 16 accepts a message input by a user and inputs the message into the large-scale language model. Then, based on the response data output from the large-scale language model, the data analysis device 16 identifies data to be analyzed from multiple types of data stored in at least one database, performs an analysis using the data to be analyzed, and outputs the analysis results. Specifically, the data analysis device 16 outputs the analysis result R1 as shown in FIG. 1 .

[0022] Also, for example, when a message such as that shown in FIG. 4, "Please create a scatter plot of X price and Y price," is entered in the input field B1, the data analysis device 16 outputs the analysis result R2 shown in FIG. 4.

[0023] Furthermore, for example, when a message such as that shown in Fig. 5, "Please provide drawing data similar to the drawing data of drawing number xxx," is input into the input field B1, the data analysis device 16 outputs the analysis result R3 shown in Fig. 5. Note that each square in the analysis result R3 represents drawing data. Since the database of this embodiment also stores drawing data, the data analysis device 16 can also output the analysis result R3 shown in Fig. 5.

[0024] 6, a file input field B3 can also be displayed on the screen S. This allows the user to drag and drop the file D1 to select the data in the file D1 as the data to be analyzed.

[0025] As shown in FIG. 7 , the screen S can also display an analysis purpose input field B4. By inputting the analysis purpose, the large-scale language model can output response data that is consistent with the analysis purpose. For example, as shown in FIG. 7 , the user can input the analysis purpose "review suppliers" in the analysis purpose input field B4 and "first calculate the order amount and number of items for each supplier" in the input field B1. In this case, analysis results R4 and R5 are output, as shown in FIG. 7 . Analysis result R4 is information that represents the analysis procedure, and selecting P in the figure displays the program code used in the analysis.

[0026] The specific details will be explained below.

[0027] (Data Analysis Device 16) The data analysis device 16 is a server for analyzing data. As shown in Fig. 1, the data analysis device 16 functionally includes a reception unit 20, a response data acquisition unit 22, an analysis unit 24, a generative model storage unit 26, a manufacturing industry document database 28, a drawing document related database 30, a drawing database 32, and a manufacturing industry structured database 34.

[0028] The generative model storage unit 26 stores a large-scale language model, which is an example of a generative model. The large-scale language model is a model that is constructed in advance using known artificial intelligence technology and known machine learning technology. The large-scale language model is also capable of processing image data.

[0029] The manufacturing document database 28 stores manufacturing document data, which is document data related to the objects displayed in the drawing data. The manufacturing document data is, for example, document files such as various specifications related to the products, catalogs for customers, or reports related to the products.

[0030] The drawing document related database 30 stores drawing document data, which is document data present in the drawing data, and drawing numerical data, which is numerical data present in the drawing data. The drawing document data is, for example, document data present in title blocks drawn in the drawing data, and includes not only characters but also processing-related symbols, etc. The drawing numerical data is, for example, numerical data such as dimensions of objects drawn in the drawing data.

[0031] The drawing database 32 stores drawing data depicting products.

[0032] The manufacturing structured database 34 stores structured data related to objects displayed in the drawing data. The structured data may include data with various attributes. For example, the structured data may be numerical data with attributes such as the dimensions, cost, unit price, number manufactured, or number sold of the object, or textual information with attributes such as the product name, material name, supplier name, or customer name.

[0033] Therefore, in this embodiment, multiple types of data related to the manufacturing industry are stored in at least one database, and the multiple types of data related to the manufacturing industry include drawing data.

[0034] (User Terminal 18) The user terminal 18 is a terminal operated by a user. Specifically, the user operates the user terminal 18 to exchange information with the data analysis device 16.

[0035] The data analysis device 16 and the user terminal 18 of the data analysis system 10 can each be realized by, for example, a computer 70 shown in FIG. 8 . The computer 70 includes a CPU 71, a memory 72 serving as a temporary storage area, and a non-volatile storage unit 73. The computer 70 also includes an input / output interface (I / F) 74 to which an input / output device (not shown) is connected, and a read / write (R / W) unit 75 that controls reading and writing of data from and to a recording medium. The computer 70 also includes a network I / F 76 that is connected to a network such as the Internet. The CPU 71, memory 72, storage unit 73, input / output I / F 74, R / W unit 75, and network I / F 76 are connected to one another via a bus 77.

[0036] The storage unit 73 can be realized by a hard disk drive (HDD), a solid state drive (SSD), a flash memory, etc. The storage unit 73 as a storage medium stores a program for operating the computer 70. The CPU 71 reads the program from the storage unit 73, loads it into the memory 72, and sequentially executes the processes contained in the program.

[0037] <Operation of Data Analysis System 10>

[0038] Next, the operation of the data analysis system 10 of this embodiment will be described.

[0039] The user operates his / her user terminal 18 to input request data for a data analysis related to the manufacturing industry, the request data being expressed in natural language (e.g., "Please create a price distribution" in FIG. 3 ), into the input field B1 on the screen S. The data analysis device 16 executes the data analysis processing routine shown in FIG.

[0040] In step S100, the receiving unit 20 receives requested data input by the user. At this time, the user may input, for example, a file D1 as shown in Fig. 6 into the screen S, thereby selecting the data D1 as the data to be analyzed.

[0041] In step S102, the response data acquisition unit 22 inputs the request data received in step S100 to the large-scale language model stored in the generative model storage unit 26. The response data acquisition unit 22 also inputs an analysis tool list, which is a list of tools used for analysis, to the large-scale language model stored in the generative model storage unit 26.

[0042] In step S104, the response data acquisition unit 22 acquires the response data output from the large-scale language model.

[0043] The response data output from the large-scale language model of this embodiment includes analysis tool data representing the type of analysis tool to be used when performing the data analysis represented by the request data, instruction data for the analysis tool, and analysis target data representing the data to be analyzed from multiple types of data related to the manufacturing industry.

[0044] For example, let us consider a case where the requested data is "10 items with surface treatment limited to black dyeing, sorted in descending order of actual price." In this case, "10 items with surface treatment limited to black dyeing, sorted in descending order of actual price" and the analysis tool list are input to the large-scale language model. A prompt to output appropriate response data is also input to the large-scale language model.

[0045] In this case, the large-scale language model outputs response data such as "Extraction to manufacturing data (argument: SELECT * ...)." Of this response data, "manufacturing data" corresponds to the data to be analyzed among multiple types of data related to the manufacturing industry. Furthermore, "extraction" of the response data corresponds to analysis tool data that indicates the type of analysis tool to be used when performing data analysis. Furthermore, "(argument: SELECT * ...)" of the response data corresponds to instruction data for the analysis tool. The instruction data is, for example, actual program code. Therefore, in the above example, the data to be analyzed is manufacturing data, the analysis tool is an extraction tool, and the specific instruction data is "SELECT * ...."

[0046] In step S106, the analysis unit 24 performs the data analysis represented by the request data received in step S100 using the analysis tool data, instruction data for the analysis tool, and data to be analyzed, which are included in the response data obtained in step S102, thereby obtaining data analysis results corresponding to the request data.

[0047] For example, the analysis unit 24 executes the instruction data, program code "SELECT * ...", using an analysis tool corresponding to the analysis tool data "extract" included in the response data, on the manufacturing structured database 34, which stores the "manufacturing data" that is the data to be analyzed. This causes a query corresponding to the request data to be executed on the manufacturing document database 28. The analysis tool corresponding to "extract" corresponds to, for example, at least one of the analysis tools listed in the analysis tool list. While the above example illustrates a case where an analysis is selected based on the word "extract," a specific analysis tool name may be output from the large-scale language model instead of the word "extract." Then, for example, the analysis unit 24 acquires a query result table: "id, price, surface treatment, ...." Next, the analysis unit 24 executes a specific analysis process using the query result table: "id, price, surface treatment, ...." For example, the analysis unit 24 acquires an analysis result corresponding to the request data by inputting the query result table "id, price, surface treatment, ...," the request data "10 results in descending order of actual price, limited to surface treatments with black dyeing," and a predetermined prompt into the large-scale language model. When performing a specific analysis, the analysis unit 24 may perform the analysis using a known analysis processing tool without using a large-scale language model.

[0048] When drawing data is identified as the data to be analyzed, the analysis unit 24 analyzes the drawing data, for example, by using a known image analysis tool. For example, when attempting to obtain the analysis result R3 shown in Fig. 5, the analysis unit 24 analyzes image data corresponding to the drawing data stored in the drawing database 32 using a known image analysis tool, and analyzes the image features in the drawing data to obtain the analysis result R3 shown in Fig. 5. Furthermore, for example, the analysis unit 24 may analyze the text data and numerical data in the drawing data using a known image analysis tool.

[0049] In step S108, the analysis unit 24 outputs the analysis results acquired in step S106, and this processing routine ends.

[0050] The analysis results output from the analysis unit 24 are sent to the user terminal 18. The user checks the analysis results displayed on the display unit (not shown) of his / her own user terminal 18.

[0051] As described above, the data analysis device according to this embodiment receives request data requesting data analysis related to the manufacturing industry, the request data being expressed in natural language. The data analysis device then inputs the received request data into a large-scale language model, which is an example of a pre-trained generative model, to obtain response data output from the large-scale language model. The data analysis device performs data analysis represented by the request data using analysis tool data representing the type of analysis tool to be used when performing the data analysis represented by the request data, instruction data for the analysis tool, and analysis target data representing the data to be analyzed among multiple types of data related to the manufacturing industry, all of which are included in the response data, to obtain data analysis results corresponding to the request data. The data analysis device then outputs the data analysis results. This allows for easy data analysis when data related to the manufacturing industry includes multiple types of data.

[0052] Furthermore, in this embodiment, at least one database stores multiple types of data related to the manufacturing industry, and the multiple types of data related to the manufacturing industry include drawing data. When drawing data is identified as the data to be analyzed, the data analysis device analyzes the drawing data by using an image analysis tool. This makes it possible to easily perform data analysis even when drawing data is included in data related to the manufacturing industry.

[0053] Furthermore, by using a data analysis device, it is possible to simultaneously analyze multiple types of data, such as images (e.g., drawing data) and documents (e.g., text in drawings or various specifications and reports), as well as numerical data such as sales or dimensions of manufactured products, making it possible for users to perform analyses related to the manufacturing industry even if they do not have specialized knowledge.

[0054] In the manufacturing industry, drawing data (image data) and sales data (numerical data) for the products depicted in the drawing are typically handled in different data formats and in different systems, methods, and locations. By using the data analysis device of this embodiment, however, it is possible to perform processes such as specifying an image and displaying sales figures through prompts. Furthermore, because user instructions can be given in natural language, data extraction processes using programming languages ​​are no longer necessary. Furthermore, the steps required for data analysis—data extraction, formatting, and visualization—can be performed with simple prompts from a single interactive screen.

[0055] The present disclosure is not limited to the above-described embodiments, and various modifications and applications are possible within the scope of the gist of the present disclosure.

[0056] For example, in the above embodiment, the data analysis device 16 has been described as storing a large-scale language model, which is an example of a generative model, in the generative model storage unit 26, which is its own storage unit. However, the present invention is not limited to this. For example, the data analysis device 16 may use a large-scale language model provided via the network 19 from an external large-scale language model system different from the data analysis device 16.

[0057] Furthermore, the large-scale language model of the above embodiment may be a model (a so-called fine-tuned model) that has been trained in advance so that when request data is input, analysis tool data, instruction data, and analysis target data are output.

[0058] Furthermore, in the above embodiment, the data analysis device 16 is described as having various databases by itself, but this is not limited to this. For example, the data analysis device 16 may acquire various data from an external database different from the data analysis device 16 via the network 19. Furthermore, in the above embodiment, the case where multiple types of data related to the manufacturing industry are stored in multiple databases is described as an example, but this is not limited to this. Multiple types of data related to the manufacturing industry may be stored in a single database.

[0059] Furthermore, various processors other than the CPU may execute the processes executed by loading software (programs) in the above embodiments. Examples of processors in this case include programmable logic devices (PLDs) (such as field-programmable gate arrays (FPGAs)) whose circuit configuration can be changed after manufacture, and dedicated electrical circuits, such as application-specific integrated circuits (ASICs), which are processors having circuit configurations designed specifically to execute specific processes. Each process may be executed by one of these various processors, or by a combination of two or more processors of the same or different types (e.g., multiple FPGAs, or a combination of a CPU and an FPGA). The hardware structure of these various processors is, more specifically, an electrical circuit that combines circuit elements such as semiconductor elements.

[0060] In the above embodiment, the programs are pre-stored (installed) in a storage device, but this is not limiting. The programs may be provided in a form stored in a storage medium such as a CD-ROM, DVD-ROM, Blu-ray Disc, or USB memory. The programs may also be downloaded from an external device via a network.

[0061] (Supplementary Notes) The following supplementary notes are provided regarding aspects of the present disclosure.

[0062] (Supplementary Note 1) A data analysis device comprising: a receiving unit that receives request data requesting data analysis related to the manufacturing industry and that is expressed in natural language; a response data acquisition unit that inputs the received request data into a pre-trained generative model and acquires response data output from the generative model; and an analysis unit that performs the data analysis represented by the request data using analysis tool data that indicates a type of analysis tool to be used when performing the data analysis represented by the request data, instruction data for the analysis tool, and analysis target data that indicates data to be analyzed among multiple types of data related to the manufacturing industry, to acquire a data analysis result corresponding to the request data, and output the data analysis result. (Supplementary Note 2) The data analysis device according to Supplementary Note 1, wherein multiple types of data related to the manufacturing industry are stored in at least one database. (Supplementary Note 3) The multiple types of data related to the manufacturing industry include drawing data, and when drawing data is identified as the analysis target data, the analysis unit analyzes the drawing data by using an image analysis tool. (Supplementary Note 4) The data analysis device according to any one of Supplementary Notes 1 to 3, wherein the multiple types of data related to the manufacturing industry include manufacturing document data which is document data related to an object displayed in the drawing data, manufacturing numerical data which is numerical data related to the object displayed in the drawing data, drawing document data which is document data present in the drawing data, and drawing numerical data which is numerical data present in the drawing data. (Supplementary Note 5) The data analysis device according to any one of Supplementary Notes 1 to 4, wherein at least a portion of the analysis target data is data input by the user. (Supplementary Note 6) The data analysis device according to any one of Supplementary Notes 1 to 5, wherein, when outputting the response data from the generative model, the response data including the analysis tool data, the instruction data, and the analysis target data is acquired by inputting, to the generative model, a prompt for outputting the analysis tool data and the instruction data.(Supplementary Note 7) The data analysis device according to Supplementary Note 6, wherein the response data acquisition unit inputs the request data and an analysis tool list, which is a list of tools used for analysis, to the generative model when outputting the response data from the generative model. (Supplementary Note 8) The data analysis device according to any one of Supplementary Notes 1 to 7, wherein the generative model is a model that has been trained in advance so that, when the request data is input, the analysis tool data, the instruction data, and the analysis target data are output. (Supplementary Note 9) A data analysis method in which a computer performs the following processes: receives request data requesting data analysis related to the manufacturing industry, the request data being expressed in natural language; inputs the received request data into a pre-trained generative model to obtain response data output from the generative model; performs the data analysis represented by the request data using analysis tool data representing a type of analysis tool to be used when performing the data analysis represented by the request data, instruction data for the analysis tool, and analysis target data representing data to be analyzed among multiple types of data related to the manufacturing industry, which are included in the response data; obtains data analysis results corresponding to the request data; and outputs the data analysis results. (Supplementary Note 10) A data analysis program for causing a computer to execute a process of receiving request data requesting data analysis related to the manufacturing industry, the request data being expressed in natural language; inputting the received request data into a pre-trained generative model to obtain response data output from the generative model; performing the data analysis represented by the request data using analysis tool data representing a type of analysis tool to be used when performing the data analysis represented by the request data, instruction data for the analysis tool, and analysis target data representing data to be analyzed among multiple types of data related to the manufacturing industry, which are included in the response data, to obtain a data analysis result corresponding to the request data, and outputting the data analysis result.(Supplementary Note 11) A data analysis system including a plurality of user terminals and a data analysis device according to any one of Supplementary Notes 1 to 8, wherein one of the plurality of user terminals transmits the request data to the data analysis device, the data analysis device receives the request data transmitted from the user terminal and transmits the data analysis results corresponding to the request data to the user terminal, and the user terminal displays the data analysis results on a display unit.

[0063] All publications, patent applications, and technical standards mentioned in this specification are herein incorporated by reference to the same extent as if each individual publication, patent application, or technical standard was specifically and individually indicated to be incorporated by reference.

Claims

1. A data analysis device comprising: a receiving unit that receives request data that requests data analysis related to the manufacturing industry and that is expressed in natural language; a response data acquisition unit that inputs the received request data into a pre-trained generative model and acquires response data output from the generative model; and an analysis unit that performs the data analysis represented by the request data using analysis tool data that indicates the type of analysis tool to be used when performing the data analysis represented by the request data, instruction data for the analysis tool, and analysis target data that indicates the data to be analyzed from multiple types of data related to the manufacturing industry, and acquires data analysis results corresponding to the request data, and outputs the data analysis results.

2. The data analysis device according to claim 1, wherein multiple types of data relating to the manufacturing industry are stored in at least one database.

3. The data analysis device according to claim 1 or claim 2, wherein the multiple types of data related to the manufacturing industry include drawing data, and when drawing data is identified as the data to be analyzed, the analysis unit analyzes the drawing data by using an image analysis tool.

4. A data analysis device as described in claim 1 or claim 2, wherein the multiple types of data related to the manufacturing industry include manufacturing document data which is document data related to objects displayed in drawing data, manufacturing numerical data which is numerical data related to objects displayed in the drawing data, drawing document data which is document data present in the drawing data, and drawing numerical data which is numerical data present in the drawing data.

5. The data analysis device according to claim 1 or 2, wherein at least a portion of the data to be analyzed is data input by a user.

6. The data analysis device of claim 1 or claim 2, wherein the response data acquisition unit, when outputting the response data from the generative model, acquires the response data including the analysis tool data, the instruction data, and the data to be analyzed by inputting the analysis tool data, the instruction data, and a prompt to output the analysis tool data, the instruction data, and the data to be analyzed into the generative model.

7. The data analysis device of claim 6, wherein the response data acquisition unit outputs the response data from the generative model by inputting the request data and an analysis tool list, which is a list of tools used for analysis, into the generative model.

8. The data analysis device according to claim 1 or claim 2, wherein the generative model is a model that has been trained in advance so that when the request data is input, the analysis tool data, the instruction data, and the analysis target data are output.

9. A data analysis method in which a computer performs the following processes: receiving request data that requests data analysis related to the manufacturing industry and that is expressed in natural language; inputting the received request data into a pre-trained generative model to obtain response data output from the generative model; performing the data analysis represented by the request data using analysis tool data that indicates the type of analysis tool to be used when performing the data analysis represented by the request data, instruction data for the analysis tool, and analysis target data that indicates the data to be analyzed from multiple types of data related to the manufacturing industry, which are included in the response data; obtaining data analysis results corresponding to the request data; and outputting the data analysis results.

10. A data analysis program for causing a computer to execute the following process: receiving request data requesting data analysis related to the manufacturing industry and expressed in natural language; inputting the received request data into a pre-trained generative model to obtain response data output from the generative model; performing the data analysis represented by the request data using analysis tool data representing the type of analysis tool to be used when performing the data analysis represented by the request data, instruction data for the analysis tool, and analysis target data representing the data to be analyzed from multiple types of data related to the manufacturing industry, which are included in the response data; obtaining data analysis results corresponding to the request data; and outputting the data analysis results.

11. A data analysis system including a plurality of user terminals and the data analysis device according to claim 1 or claim 2, wherein one of the plurality of user terminals transmits the requested data to the data analysis device, the data analysis device receives the requested data transmitted from the user terminal and transmits the data analysis results corresponding to the requested data to the user terminal, and the user terminal displays the data analysis results on a display unit.

Images