Explorer View (Dataset Level)

The Explorer view in Max enables users to query a dataset using plain English. This interface translates natural language questions into SQL, runs the query on the connected data source, and displays the results for easy interpretation.

How It Works

1. Enter a Natural Language Question
   Type your question into the Natural Language Question box. For example:
   "What were sales by month in 2022?"

2. Click Run
   Max will:

   • Interpret your question.
   • Generate SQL.
   • Execute it against the selected dataset.
   • Present the result in the Preview section.

Preview Tabs

• Data: Displays the final output of the SQL query in a table.
• Rendered Prompt: Shows how Max interpreted and structured your question.
• Explanation: Describes the generated SQL in plain English.
• SQL: Reveals the exact SQL code executed.
• Visualization: Offers a graphical representation of the results (when applicable).

Creating and Using Examples

Examples help improve SQL generation in the Data Explorer by providing the system with reference patterns for how to interpret and respond to similar questions.

Creating Examples from Data Explorer

When you find a question and answer pair that works well in Data Explorer:

1. Save as Example: Click the "Add to Examples" button
2. What gets saved:

• The natural language question you asked
• The SQL query that was generated
• The relationship between them

This example then becomes available to the Data Explorer skill in chat, helping it generate better SQL for similar questions in the future.

Managing Examples

Viewing Examples

1. Navigate to the Examples tab in Data Explorer
2. Review all currently configured examples for your dataset
3. See how questions map to SQL queries

Editing Examples

When adding an example after SQL has been generated:

1. Modify the question: You can change the natural language question before saving
2. Keep the SQL: The generated SQL remains the same
3. Save: This creates an example with your modified question paired with the original SQL

This is useful for:

• Creating variations of questions that should produce the same SQL
• Simplifying complex questions while maintaining the correct query
• Building a comprehensive example library for common query patterns

Best Practices

• Save successful queries: When Data Explorer generates correct SQL, save it as an example
• Create variations: Use the question editing feature to create multiple phrasings for the same query
• Regular review: Periodically review your examples to ensure they remain relevant
• Cover common cases: Build examples for frequently asked questions in your organization