ntelligent Copilot is an intelligent upgrade of Ask Data. Based on a new generation of AI big model, Ask Data allows users to get data analysis results from BI system through natural language session, eliminating the need for complicated production process and greatly improving the efficiency of data analysis and interactive experience.Copilot module requires Y+AIGC license. 1. Scenario
Business people or data analysts who want to do their own analysis and are not limited to the fixed dashboards provided can make use of the intelligent Copilot, which allows viewers to quickly get answers to their questions through natural language Q&A and conduct analysis freely on their own.
Users in the process of analysing data, you can directly enter the data questions to get the results, which reduces the process of data binding operations, data simple processing (such as the conversion of date expressions, the creation of four operation calculation columns, etc.), and can quickly add the results of the analysis of the dashboard, greatly improving the efficiency of the production of analysis dashboard.
2. Copilot
2.1 Function Entry
The entrance to Copilot is in the main navigation bar, as shown in the following figure.
The interface functions and descriptions are as follows.
No.
Content
Description
1
Dataset
When using Copilot, you need to first select a dataset and analyse the dataset with questions
2
Input Question
Enter the specifics of your question, which should be no more than 200 words in length. Clicking on the input box will display historical and popular questions.
3
Speech
Copilot supports voice questions, which require granting Copilot microphone privileges and clicking to ask questions by voice.
4
Data
A data fact sheet for the selected dataset, hover over each field to learn more about the data it contains.
5
Help with question
For the selected dataset, some questioning references are provided.
2.2 Functional Description
When entering Copilot, the system will load a certain dataset by default and the user can select the dataset to be analysed in the dataset selection box. In the data column on the right, hover over each field to learn more about the data it contains.
Upon entering Copilot, index information is automatically created based on the currently selected dataset so that the data is ready before the user queries it.
The index is automatically updated every 24 hours. It can also be updated manually to get the latest data information.
➢Note:
During the index creation process, the maximum load time is limited to 10s and only the first 1000 rows of data with different values are fetched for each column to ensure the performance impact on the system in case of large data volume. The maximum load time is user configurable (data.qa.create.index.max.time). When the time limit is exceeded and the index has not been fully created, only the information of the data column names will be taken for the uncreated part.
Aggregate fields on the data model do not support viewing data detail information.
2.2.1 Natural Language Questioning
As shown in the figure below, clicking on the input box will display historical and popular questions. You can select the question to ask directly or re-enter the question.
For example, if we want to know the sales of each product type, we would enter the question "What are the total sales of each product type?"
Click Go to, and Copilot intelligently analyses the graphical data based on the content of the session.
The results are charted below:
The main analytical capabilities currently supported and the corresponding sample questions are listed in the table below:
Capability
Question Example
Summary of each category
What are the total sales?
Dimension Value
What are the product categories?
Detailed Data Filter
Which orders‘ Sals more than 100？
Aggregate Data Filter
What are the provinces with total sales of more than 50,000?
Sort
Counting total sales by province and sorting them in descending order
TopN
Two provinces with the highest total sales
Specify Time
What were the sales of coffee yesterday?
Time Level
Coffee sales for the March 2016 quarter
Time Nest
Monthly coffee sales for the March 2016 quarter
Time Serie
Monthly trend in coffee sales
Year-on-Year，Month-on-Month
How much did sales increase year-on-year in January 2016
Dynamic Calculation
What is the percentage of coffee sales by province?
Note: Question descriptions should be clear and concise, avoiding ambiguity and complex descriptions. For more dialogue conversation tips and help, please refer to the Questioning Techniques and Suggestions. 2.2.2 Data description and data interpretation
In addition to returning charts, the quiz results also return chart information and an interpretation of the chart data.
The chart information contains information about the bound columns in the chart as well as filtering information to help the user quickly understand whether the chart content is consistent with expectations.
The data interpretation may contain some business advice and is provided for informational purposes only.
The data interpretation also provides a one-click copy function to quickly copy the data interpretation.
2.2.3 Good and bad
You can choose to "Good" or "Bad" the content of Copilot's answer. After clicking "Bad", "Try the following question" will be displayed, Copilot will suggest a more appropriate expression according to the content of the user's question, and the user can choose to try to change the question again if the original answer is incorrect.
2.2.4 Regeneration
If you are not satisfied with the generated result or if the result is incorrect, you can choose to regenerate. Clicking on Regenerate eliminates the current result and re-executes it.
2.2.5 Reset Conversation
If you want to end the current dialogue and clear the context, you can click Reset Session. After clicking Reset Session, Copilot will clear the contextual relationships and reopen the session.
2.2.6 Disambiguation
When conducting a session, session content and question should follow the principle of clarity and conciseness. If there is ambiguity in a question, Copilot will prompt accordingly and may pop up the session for the user to make a choice.
For example, if you want to know the sales of each product type, but the question is "What are the total sales of each product?", the system will pop up the following disambiguation prompt.
At this point select the product name and click <Confirm and Continue>. There is of course the option to re-enter the query.
Field names and field values may be selected for disambiguation, so the user can just select them according to the actual purpose.
2.2.7 Chart Operation
Save to Dashboard: Saves the current component to a dashboard, either to a new dashboard or to an existing dashboard.
Edit Copy: Click it to enter the editing page of the chart, you can edit the chart in any way you want, after editing, it supports saving to the dashboard, but it will not affect the chart in the Q&A dialogue.
2.2.8 Mobile
Copilot supports the use of App and Wap, the effect is as follows.
