Skip to main content

Uploading CSV Files to Zoë: Analysis of External Data

Reading time: 2 minutes | Difficulty: Beginner

Updated yesterday

Bring Your Own Data to Zoë

The CSV upload feature lets you analyze external data files instantly through natural conversation. No importing, no configuration - just upload and ask questions.

What You'll Learn

  • Upload single or multiple CSV files

  • Explore table structures and relationships

  • View sample data from your files

  • Identify key metrics from your columns

  • Continue analysis without re-uploading

Quick Start: Uploading Your Files

Getting Files Into Zoë

The upload process takes seconds:

  1. Click the Upload a File button at the bottom of the chat box

  2. Navigate to your folder containing CSV files

  3. Select one or multiple files (hold Ctrl/Cmd for multiple)

  4. Files appear in your chat, ready for analysis

Example: Upload four related CSV files - orders, customers, products, and inventory - all at once.

Analyzing Your Uploaded Data

Understanding Your Tables

Once uploaded, start exploring immediately:

Example Questions:

  • "Scan these tables and identify their structure"

  • "What relationships exist between these tables?"

  • "Show me all column names across the uploaded files"

Zoë automatically:

  • Maps table structures

  • Identifies potential join keys

  • Recognizes data types

  • Suggests relationships

Viewing Your Data

Get instant previews of your uploaded data:

Example Query: "Display the first 10 rows of the orders table"

Zoë will:

  • Show a formatted preview

  • Include column headers

  • Display data types

  • Highlight any data quality issues

Discovering Metrics

Let Zoë identify analytical opportunities in your data:

Example Question: "What key metrics can be built from these table columns?"

Zoë analyzes your columns and suggests:

  • Calculable KPIs

  • Possible aggregations

  • Trend analyses

  • Comparison metrics

Continuing Your Analysis

After uploading, your files remain available throughout the conversation. Build on your initial exploration:

Progressive Analysis Example:

  1. "Show me total sales from the orders file"

  2. "Break that down by product category"

  3. "Which customers contributed most to each category?"

  4. "What's the average order value by customer segment?"

Best Practices

File Preparation

  • Ensure CSV files have headers in the first row

  • Use consistent naming for related columns across files

  • Remove special characters from column names

  • Check for consistent date formats

Effective Questions

Be Specific About:

  • Table names: "in the orders table" vs "in the data"

  • Column references: "using the customer_id column"

  • Calculations needed: "sum of quantity times price"

Common Use Cases

Ad-hoc Analysis: "Analyze this vendor price list against our current costs"

Data Validation: "Compare this customer export with our database records"

External Reporting: "Create a summary of this third-party sales data"

Quick Insights: "What patterns exist in this survey response file?"

Pro Tips

Working with Multiple Files

When uploading related files:

  • Upload them together for better relationship detection

  • Ask Zoë to map connections first

  • Build analyses that span multiple tables

Memory Within Conversation

  • Uploaded files persist throughout your chat session

  • Reference them by name in subsequent questions

  • No need to re-upload for follow-up analyses

Troubleshooting

If Zoë can't read your file:

  • Verify it's a proper CSV format

  • Check for encoding issues (UTF-8 recommended)

  • Ensure file isn't corrupted or empty

If relationships aren't detected:

  • Specify join columns explicitly

  • Check for matching column names

  • Verify data type compatibility

You're Ready!

Start by uploading a CSV file you work with regularly. Ask Zoë to summarize what's in it, then explore from there. No SQL, no setup - just immediate analysis of your data.

Your first challenge: Upload a CSV and ask these three questions:

  1. "What's in this file?"

  2. "What metrics can we calculate?"

  3. "Show me the most interesting pattern in this data"


Need Help?

Related Articles:

Did this answer your question?