Agent Capabilities
The AI agent can perform a wide range of data science tasks automatically. Learn what types of analysis are available and how to get the most from each capability.
Data Analysis Capabilities
Section titled “Data Analysis Capabilities”Visual Analysis
Section titled “Visual Analysis”- Relationship visualization between variables
- Pattern recognition over time
- Distribution visualization and outlier identification
- Comparative charts across segments
Natural Language Processing
Section titled “Natural Language Processing”- Sentiment analysis of text data
- Topic extraction from unstructured content
- Content classification and categorization
- Text pattern recognition
Visual Creation
Section titled “Visual Creation”- Automatic chart selection based on data types
- Multi-variable visualizations
- Interactive dashboards
- Statistical overlays and annotations
Types of Questions You Can Ask
Section titled “Types of Questions You Can Ask”Exploratory Questions
Section titled “Exploratory Questions”- “What patterns exist in my data?”
- “Are there any outliers I should investigate?”
- “What’s the relationship between X and Y?”
- “Show me an overview of this dataset”
Comparative Analysis
Section titled “Comparative Analysis”- “How do our sales compare across regions?”
- “Which customer segments perform differently?”
- “What’s changed since last quarter?”
- “Compare performance metrics by category”
Predictive Insights
Section titled “Predictive Insights”- “What factors predict customer churn?”
- “Which variables correlate with high performance?”
- “Are there early warning signs in the data?”
- “What drives our best outcomes?”
Text Analysis
Section titled “Text Analysis”- “What are customers saying about our product?”
- “Analyze sentiment in support tickets”
- “What topics come up most in feedback?”
- “Categorize these documents by theme”
Analysis Depth
Section titled “Analysis Depth”Surface-Level Insights
Section titled “Surface-Level Insights”- Quick data summaries and overviews
- Basic statistical measures (mean, median, mode)
- Simple visualizations and charts
- High-level pattern identification
Multi-Variable Analysis
Section titled “Multi-Variable Analysis”- Multi-variable visualization with confounders
- Time-based pattern exploration
- Visual pattern detection in charts
- Group-based comparison analysis
Advanced Visualization
Section titled “Advanced Visualization”- UMAP dimensionality reduction for vector plots
- Text analysis through semantic functions
- AI-powered data transformation with PXL
- Complex filtering and data exploration
Data Types Supported
Section titled “Data Types Supported”Structured Data
- Numeric data (sales, revenue, metrics)
- Categorical data (regions, segments, types)
- Time series data (dates, timestamps)
- Boolean/binary data (yes/no, true/false)
Unstructured Data
- Text data (reviews, comments, descriptions)
- Survey responses and feedback
- Document content and reports
- Social media and communication data
Context Awareness
Section titled “Context Awareness”Session Memory
Section titled “Session Memory”- Remembers previous questions and analyses
- Builds on prior insights for follow-up questions
- Maintains context across conversation turns
- References earlier findings automatically
Data Understanding
Section titled “Data Understanding”- Automatically detects data types and patterns
- Understands relationships between columns
- Recognizes common business metrics and KPIs
- Adapts analysis approach to data characteristics
Domain Intelligence
Section titled “Domain Intelligence”- Recognizes common business scenarios
- Applies appropriate statistical methods
- Suggests relevant follow-up questions
- Provides industry-relevant insights
Performance Characteristics
Section titled “Performance Characteristics”Response Time
Section titled “Response Time”- Simple questions: 5-15 seconds
- Complex analysis: 30-60 seconds
- Large datasets: May take 1-2 minutes
- Text analysis: Varies by volume
Accuracy
Section titled “Accuracy”- Visualization rendering is precise and accurate
- Chart data reflects actual dataset values
- PXL transformations are consistently applied
- Clear methodology explanation for all charts
Scalability
Section titled “Scalability”- Handles datasets from hundreds to millions of rows
- Automatically optimizes for data size
- Provides sampling options for very large datasets
- Memory-efficient processing techniques
What’s Next?
Section titled “What’s Next?”Understanding Analysis Process
See how the agent works through problems step-by-step with transparent methodology.
Learn Best Practices
Get tips for writing effective questions and interpreting agent responses.