local GenAI CSV chatbot with anlytics using Ollama and PandasAI

whats up guys, this is a hackathon project that i worked on, GENCSV

pros:

• 100% local, no APIs required
• simple and easy to understand code
• minimal libraries required
• modular model approach (ability to choose different models from Ollama for different use cases)
• ability to improve responses using context injection

cons:

• lack of conversational skills (can't ask follow up questions)
• lack of UI
• embedding model could use some improvement (future update?)

you are only required to run and install a few packages.

STEP 1 installing OLLAMA

UNIX

Install Ollama by running the following command

curl https://ollama.ai/install.sh | sh

WINDOWS

https://ollama.com/download/windows

MAC

https://ollama.com/download/mac

STEP 2 running the Ollama server

Run the Ollama server by running the following command

ollama serve

STEP 3 choose the model you are going to use

Head down to the link below to search for the model you would like to use

https://ollama.com/library

In the example I am using the Command-r model which is specialized for RAG tasks and is also capable of conversing in 10 different languages

https://ollama.com/library/command-r https://huggingface.co/CohereForAI/c4ai-command-r-v01

STEP 4 install the libraries

pip install langchain chromadb pandasai

STEP 5 run the code

Open up the juypter notebook provided and run the cells after changing the CSV path and the model used

Analytics Generation

One feature that makes this project stand out would be the analytics generation feature