How shubham2249/data_analyst_agent works
Data Analyst Agent accepts an uploaded task/questions file, asks Gemini to generate Python, runs that code, retries on execution errors, and returns parsed list output. The README describes scraping, analysis, and visualizations; the checked-in main.py demonstrates CSV statistics and chart generation, but app.py overwrites it with generated code. Those capabilities are represented, but their runtime wiring is uncertain because generated code and its prompt-driven behavior were not inspected.
Main entry points: app.py Upload API · main.py Analysis script · task_breakdown.txt Task prompt
Components 10
API orchestration
- Upload API
app.py - Task prompt
prompts/task_breakdown.txt - List response
app.py
Generated analysis
- Analysis script
main.py - Data collection
main.py - Data analysis
main.py - Chart generation
main.py
External services
- Gemini
- Web sources
Other components
- API user
Connections 12
- API user → Upload API: uploads task
- Upload API → Task prompt: reads prompt
- Upload API → Gemini: requests code
- Gemini → Upload API: returns code
- Upload API → Analysis script: writes code
- Upload API → Analysis script: runs script
- Analysis script → Data collection: collects data
- Analysis script → Data analysis: analyzes data
- Analysis script → Chart generation: creates charts
- Data collection → Web sources: scrapes pages
- Upload API → List response: parses stdout
- List response → API user: returns list