Skip to main content

Using Code Interpreter with CSV Analysis

In this lesson, you'll learn how to feed raw CSV data to GPT-4o and analyze it using the code_interpreter tool.

Prerequisites​

Install the required packages:

pip install openai requests

The Code​

import requests
from openai import OpenAI

client = OpenAI()

# Download entire CSV as text
url = "https://raw.githubusercontent.com/wlodarzmar/csvToSqlTable/refs/heads/master/100%20Sales%20Records.csv"
print(f"Downloading CSV from {url}...")
csv_text = requests.get(url).text
print(f"Downloaded {len(csv_text)} characters")

# Compose full prompt with entire CSV embedded
input_text = f"""
You are a data analyst. Here is the complete CSV data of sales records:

\"\"\" 
{csv_text}
\"\"\"

Please write and run Python code to:

- Parse this CSV data
- Calculate total revenue (sum of 'Total Revenue' column)
- Compute average order value (average of 'Total Revenue')
- Identify the top 3 regions by total revenue
- Provide any other useful insights you find

Show your work using Python code and print the results clearly.
"""

response = client.responses.create(
    model="gpt-4o-mini",
    tools=[
        {
            "type": "code_interpreter",
            "container": {"type": "auto"}
        }
    ],
    instructions="You are a helpful analyst who can write and execute Python code.",
    input=input_text,
)

print("\n=== GPT Code Interpreter Output ===\n")
print(response.output_text)

Explanation​

  • requests.get(url).text: Downloads raw CSV file.
  • Embeds CSV directly into the prompt.
  • tools=[{"type": "code_interpreter"}]: Enables Python execution.
  • Model responds with both code and analysis in real-time.

Use Case​

Perfect for:

  • Data analysis & visualization
  • Working with CSV, JSON, and other raw data formats
  • Executable data pipelines powered by AI