“Which products are most often bundled together?"
First Analysis
Walk through creating your first analysis with Probably. You’ll learn how to load data, ask questions, and interpret the AI agent’s insights.
Step 1: Create a Dataset
Section titled “Step 1: Create a Dataset”You have three options to get started:
Upload File
Drop a CSV, Parquet, or Feather file to start analyzing immediately.
Connect Snowflake
Connect to your Snowflake data warehouse for live data analysis.
Generate Data
Use AI to create synthetic datasets for testing and exploration.
Step 2: Ask Your First Question
Section titled “Step 2: Ask Your First Question”In the query box, type:
“What factors drive higher order values?”
Step 3: Watch Probably Work
Section titled “Step 3: Watch Probably Work”You’ll see Probably systematically work through the analysis:
1
Understanding the Question
Analyzing query intent and identifying relevant variables
2
Forming Hypotheses
Generating testable hypotheses about order value drivers
3
Running Analysis
Calculating correlations, statistical significance, and creating visualizations
✓
Delivering Insights
Presenting results with evidence, visualizations, and recommendations
Step 4: Interpret Results
Section titled “Step 4: Interpret Results”Probably will present findings like:
Key Findings
- Product category explains 45% of order value variance (p < 0.001)
- Customer segment (“Premium”) averages 2.3x higher orders
- Marketing channel (“Email”) drives 31% higher values than social
- Bundle purchases increase order value by average of $127
Step 5: Ask Follow-Up Questions
Section titled “Step 5: Ask Follow-Up Questions”Based on the results, try asking:
"How can we move more customers to Premium segment?”
Understanding the Interface
Section titled “Understanding the Interface”Analysis Progress
Section titled “Analysis Progress”- Progress indicators show which step is currently running
- Real-time updates keep you informed as analysis progresses
- Cancellation option lets you stop long-running analyses
Results Display
Section titled “Results Display”- Interactive charts respond to clicks and hovers
- Statistical details available for every calculation
- Export options for charts and data
- Shareable URLs for specific findings
Follow-Up Actions
Section titled “Follow-Up Actions”- Ask related questions to dive deeper
- Switch to spreadsheet view for manual exploration
- Export findings for presentations or reports
- Save analysis for future reference
What’s Next?
Section titled “What’s Next?”Learn About Data Sources
Understand the different ways to get your data into Probably for analysis.
Explore AI Agent Capabilities
Discover the full range of questions you can ask and analyses you can perform.