How prudvireddy-git/ai-data-analyst works
AI Data Analyst is a Streamlit application for uploading CSV data (or loading it through SQLite), cleaning and exploring it, and producing charts, dataset insights, or natural-language analysis assistance. The main workflow runs through the app, which routes the dataset and user requests to specialized agents. Analyst responses are generated Pandas code and explanations; the sampled code does not execute that code. The README documents these capabilities, while implementation evidence is limited to app.py and agents.py.
Main entry points: app.py Streamlit app · agents.py Visualization agent
Components 18
User interface
- Streamlit app
app.py - Visualization request
app.py - Dataset insights
app.py - Pandas code and explanation
app.py
Dataset handling
- CSV upload
app.py - SQLite data table
- Pandas DataFrame
app.py - Cleaned dataset
app.py
Analysis agents
- Data cleaner
agents.py - Visualization agent
agents.py - Chart recommendation
agents.py - Insight agent
agents.py - Analyst agent
agents.py
Models and outputs
- Cleaned CSV download
app.py - Plotly chart
agents.py - PNG chart download
app.py - Groq LLM
Other components
- User
Connections 21
- User → Streamlit app: submits requests
- Streamlit app → CSV upload: accepts upload
- CSV upload → Pandas DataFrame: reads CSV
- Streamlit app → SQLite data table: writes and reads
- SQLite data table → Pandas DataFrame: returns rows
- Streamlit app → Data cleaner: runs cleaning
- Data cleaner → Cleaned dataset: returns cleaned data
- Streamlit app → Cleaned CSV download: exports cleaned data
- User → Visualization request: describes chart
- Streamlit app → Visualization agent: requests chart
- Visualization agent → Groq LLM: requests recommendation
- Visualization agent → Chart recommendation: parses recommendation
- Visualization agent → Plotly chart: builds chart
- Streamlit app → Plotly chart: renders chart
- Streamlit app → PNG chart download: exports chart
- Streamlit app → Insight agent: requests insights
- Insight agent → Groq LLM: requests summary
- Streamlit app → Dataset insights: displays response
- Streamlit app → Analyst agent: submits question
- Analyst agent → Groq LLM: requests analysis
- Streamlit app → Pandas code and explanation: displays response