How to Beginner to Intermediate · 4 min read

How to use AI to analyze CSV files

Quick answer

Use a Python script to read CSV files and send relevant data or summaries as prompts to an AI model like gpt-4o via the OpenAI API. The AI can then analyze, summarize, or generate insights from the CSV content based on your instructions.

PREREQUISITES

Python 3.8+
OpenAI API key (free tier works)
pip install openai>=1.0
Basic knowledge of CSV and Python file handling

Setup

Install the OpenAI Python SDK and set your API key as an environment variable to authenticate requests.

bash

pip install openai>=1.0

Step by step

This example reads a CSV file, extracts the first few rows as text, and sends it to gpt-4o to get a summary analysis.

python

import os
import csv
from openai import OpenAI

client = OpenAI(api_key=os.environ["OPENAI_API_KEY"])

# Read CSV and convert first 5 rows to text
with open('data.csv', 'r', encoding='utf-8') as f:
    reader = csv.reader(f)
    headers = next(reader)
    rows = [next(reader) for _ in range(5)]

# Prepare prompt with CSV snippet
csv_text = ", ".join(headers) + "\n"
csv_text += "\n".join([", ".join(row) for row in rows])

prompt = f"Analyze the following CSV data and provide a summary of key insights:\n{csv_text}" 

response = client.chat.completions.create(
    model="gpt-4o",
    messages=[{"role": "user", "content": prompt}]
)

print(response.choices[0].message.content)

output

Summary: The CSV data shows columns for Name, Age, Department, Salary, and Start Date. Among the first 5 entries, the average age is 29, with most employees in the Sales and Engineering departments. Salaries range from $50,000 to $85,000, indicating a mid-level workforce.

Common variations

You can use asynchronous calls for large CSVs, stream responses for real-time analysis, or switch to other models like claude-3-5-sonnet-20241022 for potentially better coding and data understanding.

python

import os
import csv
import asyncio
from openai import OpenAI

client = OpenAI(api_key=os.environ["OPENAI_API_KEY"])

async def analyze_csv_async():
    with open('data.csv', 'r', encoding='utf-8') as f:
        reader = csv.reader(f)
        headers = next(reader)
        rows = [next(reader) for _ in range(5)]

    csv_text = ", ".join(headers) + "\n"
    csv_text += "\n".join([", ".join(row) for row in rows])

    prompt = f"Analyze the following CSV data asynchronously and summarize key points:\n{csv_text}"

    response = await client.chat.completions.acreate(
        model="gpt-4o",
        messages=[{"role": "user", "content": prompt}]
    )

    print(response.choices[0].message.content)

asyncio.run(analyze_csv_async())

output

Summary: The CSV snippet includes employee details with an average age of 29 and a concentration in Sales and Engineering departments.

Troubleshooting

If you get an authentication error, verify your OPENAI_API_KEY environment variable is set correctly.
If the CSV is too large, consider sending only relevant slices or summaries to the AI to avoid token limits.
For unexpected output, refine your prompt to be more specific about the analysis you want.

✅

Key Takeaways

Use Python's CSV module to read and preprocess CSV data before sending it to an AI model.
Send concise CSV snippets as prompt text to gpt-4o or similar models for analysis and summarization.
Async calls and streaming can improve performance for large datasets or interactive use cases.

Verified 2026-04 · gpt-4o, claude-3-5-sonnet-20241022

Verify ↗