This project demonstrates a web-based application to query a dataset through natural language.

For this purpose, it uses:

• Streamlit to build a data science web app
• Pandasai to generate Pandas code from a query through OpenAI GPT-3.5

Download the dataset into the data folder at the root of the project.

If you don't have a Python environment available, you can use the conda package manager which comes with the Anaconda distribution to manage a clean Python environment.

Create a new environment and activate it:

conda create -n streamlit-pandasai python=3.9
conda activate streamlit-pandasai

Install Python dependencies in the activate Python environment:

pip install -r requirements.txt

Create a new API key and set it to the OPENAI_API_KEY environment variable beforehand.

On Windows:

set OPENAI_API_KEY="sk-..."

On Unix:

export OPENAI_API_KEY="sk-..."

Run the Streamlit project:

streamlit run streamlit_app.py