Skip to main content

Chat with Your CSV: Natural-Language Data Q&A

Level: Intermediate · Time: 40–50 min · Category: Web app

Tags: Python · Streamlit · pandas · OpenAI SDK

Build a Streamlit app where you upload a CSV and ask questions about it in plain English. The app computes a compact data profile locally with pandas and sends that profile to the Research Computing OpenAI-compatible API, which answers grounded in the profile.

Safe by design

This app does not run code the model writes. It computes a data profile with pandas and asks the model to reason over that profile. That keeps a hackathon demo safe and predictable.

Prerequisites​

Confirm Python 3 and pip are available before you start:

python3 --version      # expect 3.9 or newer
python3 -m pip --version

If Python is not found, install it from python.org. On the supercomputer, load a Python module instead (for example module load python), then run the checks again.

What you will do​

  • Create a Streamlit workspace with pandas and the OpenAI SDK.
  • Compute a compact profile of an uploaded CSV.
  • Ask questions in English and get grounded answers.
  • Add transparency and column controls.

Build it step by step​

1. Create the workspace​

mkdir csv-qa
cd csv-qa
python3 -m venv .venv
source .venv/bin/activate
pip install streamlit pandas openai python-dotenv
Keep this out of Git

If you track this project with Git, do not commit your virtual environment or any secrets. Create a .gitignore file in the project folder with at least these lines:

.gitignore
.venv/
.env
__pycache__/

.venv/ is large and specific to your machine, and .env holds your API key — neither belongs in a shared repository. New to Git? See Track a New Project with Git.

2. Add API connection values​

Your program needs three values to reach the API: your API key, the gateway URL, and a model ID. You keep them in a small text file named .env so they stay out of your code — and out of Git.

What is a .env file, and where does it go?

A .env ("dot-env") file is a plain text file that stores settings as NAME=value lines, one per line. Programs read it at startup so you never have to paste secrets like API keys directly into your code.

  • The filename is literally .env — a leading dot with nothing before it (not config.env or .env.txt).
  • It goes in your project folder: the same directory you just created, where the .venv lives and where you run your commands.
  • The python-dotenv package loads it with load_dotenv(), which looks for .env in the folder you run the program from.
  • Because it holds a secret, never commit it to Git (see the note below).

Create a new file named .env in your project folder — in VS Code use File → New File, or run code .env in the terminal — then paste these three lines and replace the placeholders:

.env
OPENAI_API_KEY="paste-your-key-here"
OPENAI_BASE_URL="https://openai.rc.asu.edu/v1"
RC_LLM_MODEL="paste-model-id-here"
Where do the values come from?

Get your API key and a valid model ID from the LLM API guide. The OPENAI_BASE_URL above is already correct for the Research Computing gateway.

  • Use the /v1 base URL, not the full /chat/completions URL, for the Python SDK.
  • Use a model ID that appears in the portal model list for your account.
  • Keep .env out of Git: add a line containing .env to your .gitignore (see Track a New Project with Git).

3. Build the profile view​

Start by showing the profile you will later send to the model. This keeps the data step separate from the API step.

You are helping me build a beginner Streamlit app that profiles a CSV.

Project goal:
- Let a user upload a CSV.
- Load it with pandas.
- Build a compact text profile that includes: row and column counts, column names with dtypes,
  df.describe(include="all") as text, missing value counts, and the first 5 rows as CSV text.
- Show the profile on the page.

Constraints:
- Keep everything in app.py.
- Add a friendly error for a malformed CSV.
- I will test with a small sample CSV.

Please:
1. Create app.py.
2. Explain how to run it.

4. Add the question box​

Now connect the profile to the API. The model answers using only the profile as context.

Please add natural-language Q&A to app.py:

Behavior:
- Add a text input for a question about the data.
- Send the question plus the profile from step 3 to the chat completions API through the OpenAI SDK.
- Read OPENAI_API_KEY, OPENAI_BASE_URL, and RC_LLM_MODEL from a .env file.
- Instruct the model to answer only from the profile, and to say clearly when the profile does
  not contain enough detail (for example, questions that need individual rows).

Constraints:
- Do not execute any code the model returns.
- Keep the app beginner-readable.
- Explain how to test it with a sample question.

5. Add column controls​

Please improve the app with one change:

- Add a multiselect so the user can choose which columns to include in the profile.

Constraints:
- Rebuild the profile from the selected columns before sending it.
- Keep the default behavior (all columns) when nothing is selected.
- Explain how to test it.

6. Add a transparency panel​

Please add one focused feature:

Feature:
- Add an expander labeled "Context sent to the model" that shows the exact profile text used for the last answer.

Constraints:
- Keep the code beginner-readable.
- Explain how this helps a user judge whether an answer is trustworthy.

More prompts to try​

  • Add a few suggested questions as buttons that fill the question box.
  • Show basic charts (histogram, value counts) for a selected column.
  • Add a note when the CSV is large explaining that the model sees a profile, not every row.

Troubleshooting​

  • If answers seem too general, remember the model only sees the profile, not the full dataset.
  • For very large files, sample rows and summarize before profiling so the context stays small.
  • If the model answers questions it cannot know from the profile, strengthen the "answer only from the profile" instruction.
  • If you get a model error, use a model ID from the portal model list for your account.

Next steps​