View User Activity Records

This article explains how to enable and view user activity records for Guandata Dashboard Insights.

Feature Overview

After the feature is enabled, the system automatically collects and stores key user actions performed in Dashboard Insights and Card Insights, such as generating insights and initiating follow-up questions. The data is synchronized to a built-in dataset, and you can also view a processed dataset that is easier to read.

Enable the Feature

Go to System Operations Page > Advanced Settings > Global Settings and enable the Front-End Tracking Data Reporting switch.

For how to access the operations page, see Introduction to the System Operations Page.

Reported Data

After the feature is enabled, front-end data is automatically reported whenever users interact with the feature. All reported data is ultimately stored in the user_use_information_record dataset.

Below is an example of data automatically reported when a user performs actions in Dashboard Insights or Card Insights, such as generating insights or asking follow-up questions:

{
  "deviceId": "xxx",
  "product": "guandata-web",
  "dataType": "BURIED_POINT_DATA",
  "time": "2025-09-26 14:11:11",
  "data": [
    {
      "type": "custom",
      "data": {
        "type": "dashboardInsight", // Operation type: dashboard or card insight
        "pgId": "Page ID", // Related dashboard ID
        "operation": "generate", // User action: generate, refresh, and so on
        "llm": "deepseek-V3" // Invoked LLM
      }
    }
  ]
}

View the Data

View the Raw Dataset (user_use_information_record)

The automatically created user_use_information_record dataset contains the full raw reporting payload.

Field Descriptions for user_use_information_record

Using Dashboard Insights events as an example:

No.	Field Name	Description	Sample / Enum Values
1	User ID	Unique identifier of the operating user	-
2	Domain	Environment domain to which the data belongs	demo
3	Device ID	Unique identifier of the device	-
4	Product	Product name that generated the data	guandata-web
5	Data Type	Type identifier of the payload; here it is tracking data	BURIED_POINT_DATA
6	Uploaded Data	Contains operation type and key action details, including insight type, action, and LLM. See Uploaded Data Field Description	{"type":"custom", "data": {"type":"dashboardInsight","pgId":"sfc62866c6bd741fbb8928d9", "operation":"generate", "llm":"Deepseek-V3"}}
7	API Request Time	Time when the interface request was made	-
8	Front-End Report Time	Time when the front-end reported the data	-
9	Trace ID	Unique identifier used for tracing the request chain	-
10	Reporter Program	Program identifier or name of the reporting process	-
11	Client OS	Operating system used by the client	Mac OS X
12	Client Browser	Browser used by the client	Chrome
13	Client IP	IP address of the client	192.168.xxx.xxx, 10.xxx.xxx.0

Uploaded Data Field Description

For Dashboard Insights-related operations, the value of Uploaded Data is a JSON string with the following structure:

Field Name	Description	Possible Values
type	Operation type. All reporting records for Dashboard Insights and Card Insights use custom	custom
data.type	Dashboard insight or card insight	dashboardInsight / cardInsight
data.pgId	Dashboard page ID. Present for both Dashboard Insights and Card Insights	sfc62866c6bd741fbb8928d9
data.cdId	Insight card ID. Present only when Card Insights is triggered	c5a2a5f879ad24f74b832293
data.operation	User action that triggered the LLM call. See data.operation Field Description	generate / refresh-data / ...
data.llm	Name of the currently invoked LLM when multiple models are available in the environment	deepseek-v3 and so on

data.operation Field Description

Scenario		data.type	data.operation
Dashboard Insights	Automatic generation when opening the sidebar	dashboardInsight	generate
	Click the `Regenerate` button		refresh-data
	Ask a follow-up question after an insight is generated		follow-up
	Launch an insight from `Custom Insight Logic`		custom-analysis
	Click `Preview` while editing Dashboard Insights configuration		preview-analysis
Card Insights	Card automatically generates insights when the page is opened	cardInsight	generate
	Click the `Re-Insight` button on the card toolbar		refresh-data
	Regeneration triggered by editing, including (1) `Preview` after modifying the insight scope or prompt (2) automatic regeneration after clicking confirm or cancel		modify-content
	Automatic re-insight when data in the selected insight scope changes, including (1) card data updates caused by filters or linked interactions (2) drill actions on cards (3) card deletion		refresh-on-source-update

View the ETL-Processed Readable Dataset

The raw dataset is relatively complex, so we provide a preset ETL job that automatically joins and processes the raw data into a more user-friendly and analysis-ready dataset.

ETL download link:Guandata Dashboard Insights Usage Processing ETL

Advantages of the processed dataset:

• Information is more intuitive: it automatically joins user name, email, page name, and card name.
• The scenario is clearer: it adds the Insight Operation Scenario field to describe actions in plain language, such as Card automatically generates insights when the page is opened.
• The data is more focused: it keeps only records related to Guandata Dashboard Insights and filters out basic actions such as page switching.

Create the ETL

1. On the Data Preparation > Smart ETL page, click Create ETL.

   

2. Click Import in the upper-right corner, upload Guandata Dashboard Insights Usage Processing ETL, click Next, and then click Confirm on the next page without making any further changes.

   
   
   
   
   

Replace the Input Datasets

After the ETL is imported, you will notice that many datasets show an exclamation mark below them, indicating that the input datasets need to be replaced.

Using user_use_information_record as an example, select user_use_information_record, click Please Select under Select Dataset on the left, search for the user_use_information_record dataset, select it, and then click Confirm.

Note

Use the same procedure to replace the remaining datasets.

Set the Output Dataset Save Path

Select Guandata Dashboard Insights Usage Record Dataset, click Please Select under Save To on the left, choose the save path, and then click Confirm.

Preview and Review the Processed Data

Select Guandata Dashboard Insights Usage Record Dataset. When the Preview icon appears at the bottom, click it to view the processed data.

After ETL processing, the generated dataset contains only usage records reported for Guandata Dashboard Insights and provides more explicit descriptions for several enum values.

The following new fields are added:

Field Name	Description	Example Value
User Name	Joined from the user ID and built-in dataset	Zhang San
email	User email address	zhangsan@guandata.com
Uploaded Data Type	The type value inside Uploaded Data	custom
Uploaded Operation Type	The data.type value inside Uploaded Data	Card Insight / Dashboard Insight
Page Name	Joined from the page ID and built-in dataset. It may be null if the draft page has not been published or the page has been deleted	Store Manager Dashboard
Card Name	Joined from the card ID and built-in dataset. It may be null if the draft page has not been published or the card has been deleted. This field exists only for Card Insights and is null for Dashboard Insights	Insight Card
Insight Operation Type	The data.operation value inside Uploaded Data	refresh-data
Insight Operation Scenario	Detailed description of data.operation	Card automatically generates insights when the page is opened