Data Agent for Analytics - Data Management - Alibaba Cloud Documentation Center

Data Agent for Analytics is a data analytics agent from the Alibaba Cloud ApsaraDB team for enterprise users. It analyzes your requirements from natural language descriptions, automatically understands your data, and suggests analysis methods. The agent then uses tools to deliver the final analysis results.

Scenarios

Business personnel: Define data requirements to quickly generate reports, such as sales trends and user behavior analysis.
Data analysts: Perform exploratory analysis. The agent can complete time-consuming tasks, such as automated feature engineering and data cleaning, and reuse Python code.
Managers: Quickly obtain deep insights and analysis of key business indicators to support data-driven business decisions.

Features

Automated data exploration and insights: Describe your needs in natural language. The agent independently understands your requirements and data, gains analytical insights, and generates reports.
Shorten the path to data value: Focus on the relationship between data results and business decisions instead of complex data processing workflows.
Seamless connection to enterprise data: Supports Alibaba Cloud ApsaraDB and data sources managed by DMS to enable immediate data analytics.

Feature limitations

This feature is currently available only in the China (Hangzhou), China (Shanghai), China (Shenzhen), China (Chengdu), China (Beijing), China (Zhangjiakou), China (Hong Kong), and Singapore regions.

Layout

Initial interface:
- Data Center: Add and upload data for analysis.
- Task Management: View the history of analysis tasks.
- User Information: Displays information about the current logon account.
- Interaction Area: The area for initial communication with the agent for a new analysis task and for displaying task execution.
Workspace interface:
- Interaction Area: The interaction area that appears after an analysis task starts.
- Task Execution Area: View the code written by the agent for each step of the analysis task and the corresponding execution results.

Procedure

Log on to the Agent console.
Choose the version that best suits your business needs. For more information about the differences between the versions, see Data Agent editions.
- Free Edition:
  On the Free version card, click Free Trial.
- Personal Edition:
  - On the Personal Edition card, click Upgrade to Personal Edition.
  - Select the region, number of seats, and subscription duration based on your needs.
    Note
    Under Usage Duration, click a monthly duration.
  - Click Buy Now and complete the purchase.
- Enterprise Edition:
  - On the Enterprise Edition card, click Upgrade to Enterprise Edition.
  - Select the region, number of Large Language Model (LLM) resource plans, and subscription duration based on your needs.
    Note
    In the Subscription Duration section, click Monthly Duration to select a subscription period from 1 to 11 months.
  - Click Buy Now and complete the purchase.
Upload a data sample. You can upload a local file or use existing data.
- Click to upload a local file. You can upload CSV, XLSX, or XLS files that are no larger than 200 MB.
- Click to use existing data. You can add existing data in the Data Center.
Describe your requirements and press Enter or click to start the analysis.
Wait for the agent to generate an execution plan. Confirm whether the plan meets your needs.
- If this meets your requirements, click Start Job.
- If the plan does not meet your needs, click Modify Task and provide more details about your requirements until the plan is satisfactory.
Wait for the plan to execute. During execution, you can view the code that the agent has written for each step and the results.
Allow the agent to create a visual web report from your data to present richer insights.

Data Center

In the navigation pane on the left, click Data Center. On the page that appears, click Add Data. Select a data ingestion method based on the data source type.

Local Upload:
Drag the file that you want to analyze to the upload area, or click the area to upload the file. After the upload is complete, click Confirm.
RDS Database, PolarDB database, AnalyticDB Database:
1. Select a Region and an Instance.
2. Enter the Database Username and Database Password.
3. Click Test link to continue.
4. Select the Database that contains the table and click Add Table.
5. Click Confirm.

Long-term memory

Long-term memory is a core feature of the agent. As you analyze data by conversing with the agent, the system automatically captures and refines key information that is valuable for future data analysis. This information is stored as the agent's long-term memory. The memory is then intelligently retrieved and applied in subsequent interactions. This process significantly improves the agent's accuracy in understanding your business needs, optimizes analysis results, and enhances the overall user experience.

Feature configuration

This feature is enabled by default to continuously optimize your analysis experience. If you do not want the system to automatically store memory, you can disable this feature by performing the following steps:

Log on to the Memory Management page.
Find the Generate memory switch and turn it off.

List description

On the Memory page, you can view and manage all stored memory. The list contains the following key information:

List Item	Description
Source	Marks the specific conversation where the memory was generated. You can click the link to go directly to the original conversation context to trace the source of the memory.
Content	Shows the key information that the system extracted and understood from the conversation.
Heat	The system calculates a "popularity score" for each memory based on how often it is retrieved and used. A higher score means the memory is more frequently used and more important for your analysis work.
Status	A memory has two statuses: • `Remembered`: Active memories in this state are retrieved and used in subsequent analysis. • `Forgotten`: Inactive memories in this state are not retrieved.
Operation	In the Actions column, you can edit and delete memories: Edit: If you find that the content of a stored memory is inaccurate or incomplete, or if business definitions change, click the edit icon for that memory. In the dialog box that appears, manually correct or complete the memory content, and then save it. This ensures that the agent learns the most accurate information. Delete: If you think a memory is outdated, invalid, or should no longer be used for analysis, click the delete icon for that memory. This action has two effects: For a `Remembered` memory: After you click delete, the memory's status changes to `Forgotten`. The record remains in the list, and you can choose to remember it again. For a `Forgotten` memory: After you click delete, the memory record is permanently removed from the list and cannot be recovered.

Data privacy statement

The agent does not use your personal data for model training or iteration. This includes uploaded files, query content, analysis results, and generated reports.
Your data is processed in an exclusive computing instance that is started on demand within your personal account. The instance is destroyed after it is released. Other accounts cannot access the data, processes, or history associated with your account.
For data that is uploaded from your local machine, the original file is stored in an isolated environment and is not visible to other accounts.
For analyses that are based on a database, the data remains in the database that you purchased and is not migrated during the computation process. After the computation is complete, all intermediate results are destroyed and are not stored.