How to Build a DataFrame Chatbot using Python: Chat With Your CSV or Excel File!

Data is the new oil, but only when refined. An AI chatbot can be your refinery.

In today’s data-driven world, we are drowning in spreadsheets — from sales reports and customer logs to academic scores and HR records. What if you could ask your data questions in plain English, and get the answers?

Welcome to the world of DataFrame Chatbots — smart assistant powered by Python, AI, and your CSV/Excel files.

In this blog you will learn how to:

• Build a chatbot that talks to your spreadsheet
• Use streamlit for a clean frontend
• Process data with pandas
• Answer natural language questions using LLM’s
• Deploy the bot locally, without any API cost

Why a DataFrame Chatbot?

The goal is to turn data into information, and information into insight.

In today’s world, data is everywhere — sales report, employee records, customer logs, survey results — usually stored as spreadsheets. While tools like Excel and Pandas are powerful, they still require:

• Technical Knowledge(e.g formulas, queries, scripting)
• Manual exploration
• Time-consuming filtering and analysis

Now Imagine this :

Instead of writing formulas or code, you just ask your spreadsheet a question — and it replies like a human

That’s the power of DataFrame Chatbot. It lets you

• Ask natural language questions about your data
• Get summary answers, insights or even charts instantly
• Do it all offline — no API, no Internet, no cloud dependency.
• Save hours of time with automated data analysis

Tools and Libraries Used:

Tool	Purpose
pandas	Handle Tabular data
streamlit	Frontend app with chat UI
langchain	Agent framework and orchestration
langchain_ollama	Interface to local LLMs via Ollama
gemma:2b (Ollama)	Small, efficient local language model

Full code to Build the Offline DataFrame Chatbot

Save as app.py

import pandas as pd
import streamlit as st
from langchain.agents import AgentType
from langchain_experimental.agents import create_pandas_dataframe_agent
from langchain_ollama import ChatOllama

# Streamlit UI setup
st.set_page_config(page_title="DataFrame Chat", page_icon="💬", layout="centered")
st.title("🤖 DataFrame ChatBot ")
st.write("Developer Prem Kumar")

# Read CSV or Excel
def read_data(file):
    if file.name.endswith(".csv"):
        return pd.read_csv(file)
    else:
        return pd.read_excel(file)

# Initialize session state
if "chat_history" not in st.session_state:
    st.session_state.chat_history = []
if "df" not in st.session_state:
    st.session_state.df = None

# File upload
uploaded_file = st.file_uploader("Choose a file", type=["csv", "xlsx", "xls"])
if uploaded_file:
    st.session_state.df = read_data(uploaded_file)
    st.write("DataFrame Preview:")
    st.dataframe(st.session_state.df.head())

# Show chat history
for message in st.session_state.chat_history:
    with st.chat_message(message["role"]):
        st.markdown(message["content"])

# User input
user_prompt = st.chat_input("Enter your prompt.✨")

if user_prompt:
    st.chat_message("user").markdown(user_prompt)
    st.session_state.chat_history.append({"role": "user", "content": user_prompt})

    # Load local LLM from Ollama
    llm = ChatOllama(model="gemma:2b", temperature=0)

    # Create agent with Pandas + LangChain
    pandas_df_agent = create_pandas_dataframe_agent(
        llm,
        st.session_state.df,
        verbose=True,
        agent_type=AgentType.OPENAI_FUNCTIONS,
        allow_dangerous_code=True
    )

    # Combine messages for agent context
    messages = [{"role": "system", "content": "You are a helpful assistant"}, *st.session_state.chat_history]

    # Get response
    response = pandas_df_agent.invoke(messages)
    assistant_response = response["output"]

    # Display and save assistant response
    st.session_state.chat_history.append({"role": "assistant", "content": assistant_response})
    with st.chat_message("assistant"):
        st.markdown(assistant_response)

Try Sample Questions

Once your file is uploaded and the app is running, ask:

• What is the total sales per region?
• Show me top 3 highest spendors?
• Which month had most transactions?
• Plot total profit by month?

AI that understands your spreadsheets is like Excel with superpowers

Real-World Use Cases

Sector	Use Case
Retail	Chat with monthly sales reprot
Education	Analyze student performance spreadsheets
HR	Interact with employee data
Finance	Summarize large accounting files
NGOs	Track impact and donations via Excel

Ready to Deploy?

You can share this app internally or deploy it locally with:

streamlit run app.py

Conclusion

With Langchain, Ollama and Stramlit, building a powerful, private AI chatbot for your spreadsheets is easier than ever — no cloud, no API keys, no limitations

• Easy to use
• Fully offline
• Highly customizable
• Secure by Design

Own your model. Own your Data. Own your AI