How wp113department-cell/data_analysis works
The product is a Streamlit CSV-analysis app. Users upload or select a dataset, ask a question, and receive a grounded answer; the first question also runs a LangGraph overview pipeline for statistics, charts, insights, and a markdown report. The UI invokes the query agent for each question and streams the overview workflow through the graph. README statements document the major result panels and Groq integration; the graph source shows node calls and parallel branches. Some UI and agent internals were not sampled, so their finer relationships are omitted.
Main entry points: main.py Streamlit app · graph.py Workflow runner · agents.py AI insights · sample_data.csv CSV dataset
Components 16
User interface
- Streamlit app
main.py - CSV dataset
data/sample_data.csv - Answer panel
main.py - Statistics panel
main.py - Charts panel
main.py - Insights panel
main.py - Report download
main.py
Analysis workflow
- Workflow runner
agent/graph.py
Analysis agents
- Question answering
agent/agents.py - Analysis planner
agent/agents.py - AI insights
agent/agents.py
Analysis results
- Statistics
agent/agents.py - Chart generation
agent/agents.py - Markdown report
agent/agents.py
Other components
- Data analyst
- Groq inference
Connections 23
- Data analyst → Streamlit app: uploads and asks
README.md - CSV dataset → Streamlit app: selects sample
README.md - Streamlit app → Question answering: asks each question
main.py - Streamlit app → Workflow runner: streams overview
main.py - Workflow runner → Analysis planner: runs planner
agent/graph.py - Workflow runner → Statistics: runs statistics
agent/graph.py - Workflow runner → Chart generation: runs visualization
agent/graph.py - Workflow runner → AI insights: runs insights
agent/graph.py - Workflow runner → Markdown report: runs report
agent/graph.py - Analysis planner → Statistics: dispatches plan
agent/graph.py - Analysis planner → Chart generation: dispatches plan
agent/graph.py - Statistics → AI insights: provides statistics
agent/graph.py - Chart generation → AI insights: joins branch
agent/graph.py - AI insights → Markdown report: provides insights
agent/graph.py - Question answering → Groq inference: uses LLM
README.md - Analysis planner → Groq inference: uses LLM
README.md - AI insights → Groq inference: uses LLM
README.md - Markdown report → Groq inference: uses LLM
README.md - Streamlit app → Answer panel: renders answer
README.md - Streamlit app → Statistics panel: renders statistics
README.md - Streamlit app → Charts panel: renders charts
README.md - Streamlit app → Insights panel: renders insights
README.md - Streamlit app → Report download: offers report
README.md