Quick Start Guide

This guide will walk you through your first analysis with Querri, from loading data to creating visualizations.

What You’ll Learn

In this quick start, you’ll:

• Start a new project
• Load data from your Library
• Ask analysis questions using natural language
• View and interpret results
• Create a visualization

Step 1: Start a New Project

Every analysis in Querri happens in a project. Projects are self-contained workspaces that store all your data, analysis steps, and results.

1. Click the New Chat or + button
2. You’ll see an empty chat interface ready for your first message

Step 2: Ask Your First Question

Querri’s AI agent responds to natural language questions. Start by describing what you want to analyze:

Examples:

• “Analyze my sales data for Q4”
• “Show me customer trends”
• “Help me understand revenue by region”

What Happens Next

When you send your first message:

1. The agent sets a goal - It translates your request into a specific analytical objective
2. Finds relevant data - It searches your Library for data sources that match your needs
3. Loads the data - It automatically loads 1-2 most relevant sources into your project
4. Begins analysis - It starts creating steps to answer your question

Be Selective with Data

The agent intentionally loads only 1-2 relevant sources to keep your analysis focused. You can always load more data later if needed.

Step 3: Understanding the Response

As the agent works, you’ll see several things happen:

Steps Appear

Each step represents a single operation:

• Source Load - Data being loaded from your Library
• Transformation - SQL queries filtering or aggregating data
• Visualization - Charts and graphs
• Analysis - Statistical calculations or forecasts

Steps execute in order, with each building on previous results.

QDF Tables

When steps produce tabular data, you’ll see it in the QDF (Querri Data Frame) viewer:

• Interactive scrolling and exploration
• Automatic type detection
• Sortable columns
• Download options

Agent Messages

The agent explains what it’s doing and asks clarifying questions when needed:

• “Which date column should I use?”
• “Did you want monthly or daily aggregation?”

Answer these to help the agent proceed correctly.

Step 4: Follow-Up Questions

The real power of Querri is the conversation. Build on your analysis by asking follow-up questions:

Building on Previous Steps:

• “Now filter that to just premium customers”
• “Create a chart from those results”
• “Show me the top 10”

Exploring Different Angles:

• “What if we look at it by region instead?”
• “How does this compare to last year?”
• “Break that down by product category”

Modifying Visualizations:

• “Make that a line chart instead”
• “Sort by revenue descending”
• “Add a trend line”

The agent maintains context from earlier in the conversation, so you can refine your analysis naturally.

Step 5: Creating Visualizations

To create a chart, simply ask for it:

"Create a bar chart of sales by region"
"Show monthly trends as a line chart"
"Plot revenue vs. marketing spend"

The agent will:

1. Ensure data is in the right format
2. Select the appropriate chart type
3. Apply Querri’s color scheme and styling
4. Generate an interactive Plotly chart

Chart Best Practices

For the best results:

• Be specific about what to plot: “line chart of revenue by month”
• Mention groupings if relevant: “grouped by product category”
• Request modifications after viewing: “add a trend line” or “use log scale”

Step 6: Examining Data

If you want to investigate data more deeply, just ask:

"What's in the sales_data step?"
"Show me statistics for the revenue column"
"What are the unique values in the category column?"

The agent will run queries and provide detailed insights about your data structure and contents.

Step 7: Creating a Dashboard

Once you’ve created several useful visualizations, turn them into a dashboard:

"Create a dashboard from these charts"
"Make a dashboard with the sales and revenue charts"

The agent will:

• Analyze your project steps
• Suggest an optimal layout
• Create a dashboard with 4-12 widgets
• Set up appropriate sizing and positioning

You can then share the dashboard with your team or keep it for personal use.

Common Patterns

Pattern 1: Explore → Filter → Visualize

You: "Analyze my customer data"
Agent: [Loads customer data, shows preview]

You: "Filter to customers from 2024"
Agent: [Creates filtered dataset]

You: "Create a chart showing customer growth by month"
Agent: [Generates line chart]

Pattern 2: Compare Multiple Metrics

You: "Show me revenue and costs over time"
Agent: [Creates aggregated dataset with both metrics]

You: "Plot them on the same chart"
Agent: [Generates dual-axis line chart]

You: "Add profit margin as a third line"
Agent: [Updates chart with calculated margin]

Pattern 3: Forecast Future Trends

You: "Show me monthly sales for 2024"
Agent: [Aggregates and displays data]

You: "Forecast the next 3 months"
Agent: [Runs time series forecast, shows predictions with confidence intervals]

You: "Create a chart with historical and predicted values"
Agent: [Generates chart with dashed prediction line]

Tips for Success

1. Start Simple

Begin with a basic question and build complexity through conversation:

• ✅ “Analyze sales data” → “Filter to 2024” → “Show monthly trends”
• ❌ “Filter 2024 sales, aggregate by month, create chart, and forecast Q1 2025”

2. Check Results Along the Way

Review each step’s output before moving forward. This helps catch issues early.

3. Be Conversational

You don’t need formal language or SQL syntax:

• ✅ “Show me what’s going on with revenue this year”
• ❌ “SELECT revenue FROM sales WHERE year = 2024”

4. Use the Agent’s Suggestions

The agent often offers helpful alternatives:

• “I can also show you growth rates if that’s helpful”
• “Would you like me to break this down by customer segment?“

5. Reference the Library

If you know the name of a data source in your Library, mention it:

• “Load the Q4 Sales Report”
• “Use the customer_database source”

What’s Next?

Now that you’ve completed your first analysis, explore:

• Core Concepts - Understand how Querri works
• Prompting Guide - Learn to write effective questions
• Loading Data - Master data management
• Analysis Capabilities - Discover what’s possible

Need Help?

• Check the FAQ for common questions
• Review Agent Tools Reference to see what’s possible
• Experiment! The agent is designed to handle natural conversation